Hacking Vim
A cookbook to get the most out of the latest Vim editor
From personalizing Vim to productivity optimizations: Recipes to make life easier for experienced Vim users
Kim Schulz
BIRMINGHAM - MUMBAI
Hacking Vim A cookbook to get the most out of the latest Vim editor Copyright © 2007 Packt Publishing
All rights reserved. No part of this book may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, without the prior written permission of the publisher, except in the case of brief quotations embedded in critical articles or reviews. Every effort has been made in the preparation of this book to ensure the accuracy of the information presented. However, the information contained in this book is sold without warranty, either express or implied. Neither the author, Packt Publishing, nor its dealers or distributors will be held liable for any damages caused or alleged to be caused directly or indirectly by this book. Packt Publishing has endeavored to provide trademark information about all the companies and products mentioned in this book by the appropriate use of capitals. However, Packt Publishing cannot guarantee the accuracy of this information.
First published: May 2007
Production Reference: 1140507
Published by Packt Publishing Ltd. 32 Lincoln Road Olton Birmingham, B27 6PA, UK. ISBN 978-1-847190-93-2 www.packtpub.com
Cover Image by www.visionwt.com
Credits Author Kim Schulz Reviewers Brian Jørgensen
Project Manager Patricia Weir Project Coordinator Sagara Naik
James Eaton-Lee Kenneth Geisshirt Sven Guckes Development Editor Nanda Nag
Indexer Bhushan Pangaonkar Proofreader Chris Smith
Nikhil Bangera Layouts and Illustrations Technical Editor
Shantanu Zagade
Ajay S Cover Designer Editorial Manager Dipali Chittar
Shantanu Zagade
About the Author Kim Schulz has an M.Sc. in Software Engineering from Aalborg University in Denmark. He has been an active developer in the Linux and Open Source communities since 1997 and has worked with everything from translation and bug fixing to producing full-blown software systems. This entire time, Vim has been Kim's editor of choice and it has been the first thing he installs whenever he sits at a new computer. Today Kim works as a full-time software engineer at CSR Plc. developing software for the next generation wireless technologies. A lot of Kim's spare time has been spent on developing the open-source CMS Fundanemt. This has lead to him now owning the web-hosting company Devteam Denmark that specializes in hosting and development of Fundanemt-based websites. I would like to thank my girlfriend, Line, for letting me take the time to write this book. Without her positive attitude and help, I would never have got this book ready. I would also like to add a great thank you to Bram Moolenaar, for developing the Vim editor, and making it what it is today.
About the Reviewers Brian Jørgensen was born in 1982 in Northern Jutland, Denmark. His early
interest in computers and programming on the Commodore 64 and later the PC resulted in him studying computer science at Aalborg University, from where he is about to graduate with a masters degree in software engineering. In the late 90s he found a great interest in the Linux operating system, which he has been using since. When he is not studying or working as a freelance developer, his spare time is spent working on Open Source software projects. He is one of the core developers on the Fundanemt CMS. His main interests are in programming languages, Internet technologies, and keeping a tap on a wide range Open Source projects. Brian has a personal website (http://qte.dk/) where he runs a blog and writes about his software projects.
James Eaton-Lee works as a Consultant specializing in Infrastructure Security and has worked with clients ranging from small businesses with a handful of employees to multinational banks. He has a varied background, including experience working with IT in ISPs, manufacturing firms, and call centers. James has been involved in the integration of a range of systems, from analogue and VOIP telephony to NT and AD domains in mission-critical environments with thousands of hosts, as well as Unix and Linux servers in a variety of roles. James is a strong advocate of the use of appropriate technology, and the need to make technology more approachable and flexible for businesses of all sizes, but especially in the SME marketplace in which technology is often forgotten and avoided. James has been a strong believer in the relevancy and merit of Open Source and Free Software for a number of years and —wherever appropriate—uses it for himself and his clients, integrating it fluidly with other technologies.
Kenneth Geisshirt is a chemist by education, and is a strong free-software
advocate. He spent his Christmas holidays in 1992 installing SLS Linux, and GNU/Linux has been his favorite operating system ever since. Currently, he does consultancy work in areas like scientific computing and Linux clusters. He lives in Copenhagen, Denmark with his partner and their two children. You can find him at http://kenneth.geisshirt.dk/.
Table of Contents Preface Chapter 1: Introduction Vi, Vim, and Friends vi STEVIE Elvis Nvi Vim Vile Compatibility Vim is Charityware Summary
Chapter 2: Personalizing Vim
Where are the Config Files? vimrc gvimrc exrc Changing the Fonts Changing Color Scheme Personal Highlighting A More Informative Status Line Toggle Menu and Toolbar Adding Your Own Menu and Toolbar Buttons Adding a Menu Adding Toolbar Icons Modifying Tabs Work Area Personalization Adding a More Visual Cursor
1 7
8 8 9 9 10 11 12 13 14 14
15
16 16 17 17 18 19 20 24 25 26 27 30 30 34 35
Table of Contents
Adding Line Numbers Spell Checking Your Language Adding Helpful Tool Tips Using Abbreviations Modifying Key Bindings Summary
36 37 40 43 45 47
Chapter 3: Better Navigation
49
Chapter 4: Production Boosters
69
Faster Navigation in a File Context-Aware Navigation Navigating Long Lines Faster Navigation in Vim Help Faster Navigation in Multiple Buffers Open Referenced Files Faster Search and You Will Find Search the Current File Search in Multiple Files Search the Help System X Marks the Spot Visible Markers—Using Signs Hidden Markers—Using Marks Summary Using Templates Using Template Files Abbreviations as Templates Using Tag Lists Easier Taglist Navigation Other Usages of Taglists Using Auto-Completion Auto-Completion with Known Words Auto-Completion using Dictionary Lookup Omni-Completion All-in-One Completion Using Macro Recording Using Sessions Simple Session Usage Satisfy your own Session Needs
Sessions as a Project Manager Registers and Undo Branching Using Registers
50 50 55 55 57 58 59 59 61 62 63 64 66 68 70 70 72 74 77 77 78 79 80 81 83 85 87 87
90
91 92 93
The Unnamed Register
94
[ ii ]
Table of Contents The Small Delete Register The Numbered Registers The Named Registers The Read-Only Registers The Selection and Drop Registers The Black Hole Register Search Pattern Register The Expression Register
94 94 95 95 95 96 96 96
Using Undo Branching Folding
97 100
Using vimdiff to Track the Changes Navigation in vimdiff Using Diff to Track Changes Open Files Anywhere Faster Remote File Editing Summary
105 106 108 108 110 111
Chapter 5: Advanced Formatting
113
Simple Text File Outlining
Formatting Text Putting Text into Paragraphs Aligning Text Marking Headlines Creating Lists Formatting Code Autoindent Smartindent Cindent Indentexpr Fast Code Block Formatting
104
113 113 116 117 119 120
121 122 122 123 123
Auto Format Pasted Code Using External Formatting Tools
126 127
Summary
130
Indent Berkeley Par Tidy
127 128 129
Chapter 6: Vim Scripting
Syntax-Color Schemes Your First Syntax-Color File Syntax Regions Color Scheme and Syntax Coloring Using Scripts Script Types Installing Scripts [ iii ]
133
133 134 135 139 140 140 141
Table of Contents Uninstalling Scripts
142
Script Development Scripting Basics
143 144
Script Structure Scripting Tips
164 170
Debugging Vim Scripts Distributing Vim Scripts Making Vimballs
173 177 177
Using External Interpreters
181
Types Variables Conditions Lists and Dictionaries Loops Creating Functions
144 146 149 152 156 160
Gvim or Vim? Which Operating System? Which Version of Vim? Printing Longer Lines
170 171 171 173
Remember the Documentation
178
Vim Scripting in Perl Vim Scripting in Python Vim Scripting in Ruby
182 184 185
Summary
187
Appendix A: Vim Can Do Everything
191
Appendix B: Vim Configuration Alternatives
203
Index
211
Vim Games Game of Life Nibbles Rubik's Cube Tic-Tac-Toe Mines Sokoban Tetris Programmers IDE Mail Program Chat with Vim
Tips for Keeping your Vimrc Clean A Vimrc Setup System Storing Vimrc Online
[ iv ]
191 191 192 193 193 194 195 196 196 199 200 203 205 209
Preface Back in the early days of the computer revolution, system resources were limited and developers had to figure out new ways to optimize their applications. This was also the case with the text editors of that time. One of the most popular editors of that time was an editor called Vim. It was optimized to near-perfection for the limited system resources on which it ran. The world has come a long way since then, and even though the system resources have grown, many still stick with the Vim editor. At first sight, the Vim editor might not look like much. However, if you look beneath the simple user-interface, you will discover why this editor is still the favorite editor for so many people, even today! This editor has nearly every feature you would ever want, and if it's not in the editor, it is possible to add it by creating plugins and scripts. This high level of flexibility makes it ideal for many purposes, and it is also why Vim is still one of the most advanced editors. New users join the Vim user community every day and want to use this editor in their daily work, and even though Vim sometimes can be complex to use, they still favor it above other editors. This is a book for these Vim users. With this book, Vim users can make their daily work in the editor more comfortable and thereby optimize their productivity. In this way they will not only have an optimized editor, but also an optimized work-flow. The book will help them move from just using Vim as a simple text editor to a situation where they feel at home and can use it for many of their daily tasks. Good luck and happy reading!
Preface
What This Book Covers
Chapter 1 introduces Vim and a few well-known relatives; their history and relation to vi is briefly described. Chapter 2 introduces how to make Vim a better editor for you by modifying it for your personal needs. It shows you ways of modifying fonts, the color scheme, the status line, menus, and toolbar. Chapter 3 introduces some of the ways in which Vim helps us to navigate through files easily. It explains an alternative way for boosting navigation through files and buffers in Vim. Chapter 4 introduces you to features in Vim. It describes how to use templates, auto-completion, folding, sessions, and working with registers. Chapter 5 introduces simple tricks to format text and code. It also discusses how external tool can be used to give Vim just that extra edge it needs to be the perfect editor. Chapter 6 is especially for those who want to learn how to extend Vim with scripts. The chapter introduces scripting basics, how to use and install/uninstall scripts, debugging scripts, and lots more. Appendix A has a listing of games that have been implemented with Vim scripting; it also provides an overview of chat and mail scripts and has a section on using Vim as an IDE. Appendix B shows how to keep your Vim configuration files well organized and retain your Vim configuration across computers by storing a copy of it online
What You Need for This Book
Over the course of the last decade, Vim has evolved into a feature-rich editor. This means that the some of the features from the later versions of Vim are not accessible in the older versions of Vim. Vim is available for a wide variety of platforms and not all recipes might work on all platforms. This is typically due to the use of system-specific functionality that is not available on other platforms. This book will focus on two of the platforms where Vim is most widespread, namely Linux and Microsoft Windows. As the Linux system resembles the system used in most Unix platforms, the recipes will work onother *NIX platforms.
[]
Preface
You can find the latest source code and binary packages for the Vim Editor at www.vim.org. If you use Linux it is, however, most likely that Vim is already packed with your Linux distribution as it is the default editor on most Linux systems.
Conventions
In this book, you will find a number of styles of text that distinguish between different kinds of information. Here are some examples of these styles, and an explanation of their meaning. There are three styles for code. Code words in text are shown as follows: "We can include other contexts through the use of the include directive." A block of code will be set as follows: :amenu :amenu :amenu :amenu
Tabs.&Delete :confirm tabclose Tabs.&Alternate :confirm tabn # <silent> Tabs.&Next :tabnext <silent>Tabs.&Previous :tabprevious
Any command-line input and output is written as follows: :amenu icon=/path/to/icon/myicon.png ToolBar.Bufferlist :buffers
New terms and important words are introduced in a bold-type font. Words that you see on the screen, in menus or dialog boxes for example, appear in our text like this: "clicking the Next button moves you to the next screen".
Warnings or important notes appear in a box like this.
Tips and tricks appear like this.
[]
Preface
This book primarily focuses on the features available in Vim version 7+, but since some of the recipes cover tasks available in earlier versions, each recipe will be marked with one of the following icons that specify the version for which the recipe is applicable:
Reader Feedback
Feedback from our readers is always welcome. Let us know what you think about this book, what you liked or may have disliked. Reader feedback is important for us to develop titles that you really get the most out of. To send us general feedback, simply drop an email to
[email protected], making sure to mention the book title in the subject of your message. If there is a book that you need and would like to see us publish, please send us a note in the SUGGEST A TITLE form on www.packtpub.com or email
[email protected]. If there is a topic that you have expertise in and you are interested in either writing or contributing to a book, see our author guide on www.packtpub.com/authors.
Customer Support
Now that you are the proud owner of a Packt book, we have a number of things to help you to get the most from your purchase.
Downloading the Example Code for the Book
Visit http://www.packtpub.com/support, and select this book from the list of titles to download any example code or extra resources for this book. The files available for download will then be displayed. The downloadable files contain instructions on how to use them.
[]
Preface
Errata
Although we have taken every care to ensure the accuracy of our contents, mistakes do happen. If you find a mistake in one of our books—maybe a mistake in text or code—we would be grateful if you would report this to us. By doing this you can save other readers from frustration, and help to improve subsequent versions of this book. If you find any errata, report them by visiting http://www.packtpub. com/support, selecting your book, clicking on the Submit Errata link, and entering the details of your errata. Once your errata are verified, your submission will be accepted and the errata added to the list of existing errata. The existing errata can be viewed by selecting your title from http://www.packtpub.com/support.
Questions
You can contact us at
[email protected] if you are having a problem with some aspect of the book, and we will do our best to address it.
[]
Introduction The Vim editor (or Vi IMproved) was first released by Bram Moolenaar in November 1991 as a clone of the Unix editor vi for the Amiga platform. The first release of Vim for the Unix platform was out a year later and right away, it started to become an alternative to the vi editor. The combination of a more liberal licensing model and the fact that Vim started to become a superset of vi's functionality resulted in it becoming progressively more popular with the Open Source community. Soon more and more Linux distributions started to adopt Vim as an alternative to vi. Even if the users thought they used vi (if they actually executed the vi command) they opened Vim (the vi command had simply been substituted with a link to the vim command, which has often lead to the misunderstanding that vi and Vim are actually the same program). During the late 90s, Vim took over where vi was lacking behind in the so-called editor-war that existed between the vi editor and the Emacs editor. Bram implemented a lot of the missing features that the Emacs community used as arguments for why Emacs was better than vi/Vim, but he did it without ever neglecting the main focus areas that the vi editor had had, right from the beginning. Today, Vim is a feature-rich, fully configurable editor loved by many. It supports syntax-highlighting of more than 200 different programming languages, auto-completion for a fast growing number of languages, folding, undo/redo, multiple buffers/windows/tabs, and a lot of other features.
Introduction
Vi, Vim, and Friends
Vim is just one of many derivatives of the original vi that Bill Joy released back in 1976. Some have a feature list very close to that of vi, while others have chosen to add a wide variety of new features. Vim belongs to the group of vi clones that has chosen to add extra features. In the next section, we will introduce some of the better-known clones of vi and briefly describe the distinct features that each clone has.
vi
Vi is the original root of the Vim family tree. It was created by Bill Joy in 1976 for one of the earlier versions of BSD (Berkeley Software Distribution). The editor was an extension of the most common editor at that time, ex. Ex was, in turn, an extension of the Unix editor 'ed'. The name 'vi' is actually an abbreviation of 'visual in ex'. As the name indicates, vi was actually just a command that started the ex editor in one of its modes—the visual mode. Vi was one of the first editors to introduce the concept of modality. What this means is that the editor has different modes for different tasks—one mode for editing text, another for selecting text, and yet another for executing commands. This modality is one of the main features in vi that makes enthusiasts like the editor, but it is also what makes others dislike it even more. Not much has changed in vi since the first version, but it is still one of the most used editors in the Unix community. This is mainly because vi is considered a required application for a Unix to comply with the Single Unix Specification (SUS)—and hereby be able to call itself a Unix. []
Chapter 1
STEVIE
In 1987, Tim Thompson got his first Atari ST. In this platform, there weren't any really good editors so he decided to clone the editor vi, which was known from the Unix platform. In June 1987, he released an editor under a license that resembles what has later become known as open-source. He released it on Usenet and named it STEVIE—an abbreviation for 'ST Editor for VI Enthusiasts'. It was very simple and only provided a very small subset of the functionality that vi provided. It did, however, provide a familiar environment for vi users moving to the ST. After the release, Tim Thompson discontinued work on the editor. But soon Tony Andrews took over, and within a year he had ported it to Unix and OS/2. More features were added along the way but at some point around 1990, the development stopped. STEVIE as an editor might not have survived throughout the years, but since both Tim and Tony released the source code on Usenet as public-domain for anyone to use, a lot of the later vi clones have been both inspired and based on this code.
Elvis
STEVIE was one of the more common editors around. It was, however, full of bugs and had some quite unpractical limitations. Steve Kirkendall, who at that time used the operating system Minix, noticed one very big limitation, i.e. the STEVIE editor held the entire file in memory while editing. This was not an optimal solution when using Minix, so Steven decided to rewrite the editor to use a file as buffer instead of editing in RAM. This turned into Elvis, version 1.0. Even though Elvis was an improvement over the vi editor, it still suffered from some of the same limitations that vi had—max length of lines and only a single file buffer. Steve Kirkendall decided to rewrite Elvis completely to get rid of the limitations, and this turned into Elvis version 2, which is the generation of the editor currently available (version 2.2). With generation 2 of Elvis, Steve also included support for a range of other features that weren't in the original vi editor. Among these, a few features that are interesting and worth mentioning are: • • • •
Syntax highlighting Multiple windows support Networking support (HTTP and FTP) Simple GUI front ends []
Introduction
Elvis is not actively developed anymore, but is still widely used. It is available for Unix, MS Windows (console or GUI with WinElvis), and OS/2. The latest version of the Elvis editor can always be found here: http://elvis.the-little-red-haired-girl.org/
Nvi
Nvi, or new vi (as its full name is) is a result of a license dispute between AT&T and the Computer Science Research Group (CSRG) at University of California, Berkeley. Vi was based on an original code from the editor ed, which was under the AT&T System V Unix license, so it was not possible for CSRG to distribute vi with BSD. CSRG decided to replace the vi editor with an alternative editor under a freer license—their own BSD license. Keith Bostic was the man that took on the job to make the new vi. The vi clone Elvis was already freely available, but Keith wanted an editor that resembled the original vi editor even more. He took the code for Elvis and transformed it into an almost 100% vi compatible clone—the nvi editor. Only the Open Mode and the lisp edit option from the original vi functionality set is left out. By the release of 4.4BSD, the vi editor was completely substituted by nvi, and the software distribution was once again completely covered by a free license. Today nvi is the default vi editor in most BSD-derived distributions like NetBSD, FreeBSD, and OpenBSD, and has evolved into a more feature-rich editor than the original vi. Compared to the original vi editor, nvi has been extended to support new features like: •
Multiple edit buffers
•
Unlimited Undo
•
Extended Regular Expressions
•
CScope support
•
Primitive scripting support in Perl and Tcl/Tk
Keith Bostic is still the maintainer of the nvi source code, but not much development has been done to the code for some time now.
[ 10 ]
Chapter 1
The latest version of the nvi editor can always be found here: http://www.bostic.com/vi/
Vim
The editor Vim is the golden child of the vi family. Ever since Bram Moolenaar released the first version of Vim to the public in November 1991, this editor has evolved into one of the most feature-rich editors around. The first version of Vim was, like the Elvis editor, based on the source code of the editor Stevie. Bram, however, released Vim only for the Amiga platform, which was one of the most widespread platforms, at that time, among home computer enthusiasts. At that time Vim was an abbeviation for Vi-IMitation, which described Vim quite well in that it simply tried to do what vi did. A year later, in 1992, however, Bram made a port of his Vim editor for the Unix platform. The result of this was that Vim went beyond simply being a clone of vi for a different platform, to becoming a competitor. The development of Vim was quick and fast, and soon Vim had a range of features that the original vi editor did not have. Because of this, the abbreviation Vim was at some point changed into being Vi-IMproved instead of Vi-IMitation. Within a couple of years, Vim grew to having a range of features that a lot of vi users missed. This made more and more users switch over to using Vim instead of vi as their primary editor. In 1998, the fifth generation of Vim was released, and with it one of the most used features of today, scripting, was introduced. Now, it was possible for the user to write their own scripts for Vim, and in that way expand the functionality of the editor. This was a really strong addition to the feature set of Vim, because it would normally have required coding in a lower-level language and recompilation of the editor in order to add even simple features. A lot of features have been added to Vim throughout the last decade, and many of these are quite unique compared to the other editors and vi clones in particular. Here we will list just a few of the more distinct features of Vim, since the complete feature list would be too long: •
Editing multiple files in multiple buffers, windows, and tabs
•
Advanced Scripting language
•
Support for scripting in Perl and Python [ 11 ]
Introduction
•
Syntax highlighting for 200+ programming languages
•
Unlimited undo/redo with branching
•
Context-aware completion of words and functions
•
Advanced pattern-matching with Regular Expressions
•
Close integration with a wide range of compilers, interpreters, and debuggers
•
More than 1500 Vim scripts freely available online
Vim is available for an enormous variety of platforms like all types of Unix, Linux, MS Dos, MS Windows, AmigaOS, Atari MiNT, OS/2, OS/390, MacOS, OpenVMS, RISC OS, and QNX.
Vile
Vile is maybe the vi clone that looks least like the original vi editor—some would even say that it's not a clone at all. Vile is actually an attempt to bring the best of two worlds together in one editor: the modality of vi and the feature set of Emacs. This also explains the name Vile, which is short for "VI Like Emacs." The Vile editor project was started by Paul Fox during the summer of 1990. The code was based on the core code from the public-domain editor MicroEmacs. Paul then modified it to have modality and other vi-like features. The MicroEmacs code did not have all the features of the Emacs editor, but it had support for long lines and editing multiple files in multiple windows at the same time. These were features that vi did not have and which many programmers needed in their editor. A lot of work was done to get the MicroEmacs code to be more vi-like, and several other developers joined the project. Thomas E. Dickey joined the project in 1992 and added a wide variety of features to Vile and fixed a lot of bugs in the code. In 1994, Kevin Buettner joined the project and started working on the GUI version of vile— xvile. He added support for some of the most common widget sets at that time, like Athena, OpenLook, Motif, and the Xt Toolkit. Today Thomas is the primary maintainer of Vile and the development is steered by him. His time for working on the editor is, however, very limited. So, it is mostly only bugfixes that he adds to the editor.
[ 12 ]
Chapter 1
Vi and Vile are not very similar in the way they work, and only a minor subset of the vi features are present in Vile. The main feratures of Vile are: •
Editing modes—one mode for each file type
•
Vile procedure language for macros
•
(Experimental) Perl Support
•
Named functions that can be bound to keys as the user wishes
Vile is available for Unix, Linux, BeOS, OS/2, VMS, and MS Windows and exists in both a console version and a GUI version. The latest version of the vile editor can always be found here: http://www.vile.cx/
Compatibility
Though all the vi clones have at some point tried to behave like the vi editor, most of them have evolved in very different directions. This means that even though a lot of them support features such as syntax highlighting, they do not necessarily implement them in the same way. Therefore A syntax file from Vim cannot be used in Elvis. Even the features that originate from vi are not necessarily implemented the same way. Some of the clones have implemented features less accurately than others. Maybe the idea behind the feature is the same, but the actual result of using it is completely different. In the following table, I have tried to give a percentage of how accurately the mentioned clones resemble the vi editor (0% being least compatible and 100% being completely compatible). The comparison has been done by looking at how much effort the clone developers have made in order to implement the features of vi as precisely as possible. Clone
vi compatibility
Comment
STEVIE
10%
Only a very small feature set in common.
Vile
10%
Only general concepts like modes in common.
Elvis
80%
Large feature set in common, some features behave quite differently though.
Nvi
95%
Nearly perfect compatibility, but a range of the features behave differently.
Vim
99%
In the 'compatible mode' nearly all features are compatible. [ 13 ]
Introduction
In the table, only the features that the clones share with vi are considered. This means that even though for example, Vim has a lot of features that vi does not have, it still resembles vi very precisely on the features that they share. Besides this, Vim implements nearly all of the features that vi has. Only some of the features that Bram Moolenaar considered as bugs in vi are implemented differently in Vim. Note that in order to make Vim 99% compatible with vi, you will have to set it into compatible mode with the command: :set compatible
In Vim you can read more about vi and Vim differences with the command: :help vi-differences.
Another interesting observation is that even though STEVIE implemented a subset of the vi functionality very accurately, it did not implement enough of the vi features to be considered a close relative.
Vim is Charityware
Bram Moolenaar, the developer of the Vim editor, has chosen to release Vim under a so-called charityware license. What this means is that you can copy Vim as much as you like, but in exchange you are encourage to make donations to a charity. You can read more about the project if you open Vim and execute the command: :help uganda
You can also get more information about how you can sponsor the Vim project if you go to the website http://www.vim.org/sponsor/. As a Vim sponsor, you will get to vote for new features that should be implemented in Vim. So besides supporting a good cause, you will also get some say on how Vim will evolve in the future.
Summary
In this chapter, we introduced Vim and looked at what this book is about. Vim is just one of many clones of the old Unix editor vi, so to get a broader view of the vi-family tree, this chapter introduced some of the more well-known clones. Their history and relation to vi were briefly described and we learned that even though the vi clones at some point have tried to be like vi, they are not really compatible with each other. [ 14 ]
Personalizing Vim If you tend to use your computer a lot for editing files, you soon realize that having a good editor is of paramount importance. A good editor will be your best friend and help you with your daily tasks. But what makes an editor good? Looking at the different editors available, we see that some of them try to be the best editor by developers adding features they think the users need. Others have accepted that they are not the best editor and instead try to be the simplest most, user-friendly, or fastest-loading editor around. With the Vim editor, no one has decided what's best for you. Instead you are given the opportunity to modify a large range of settings to make Vim fit your needs. This means that the power is in the hands of the user, rather than the hands of the developers of the editor. Some settings have to do with the actual layout of Vim (e.g. colors and menus), while others change areas that affect how we work with Vim—like key bindings that map certain key combinations to specific tasks. In this chapter we will introduce a list of recipes that will help you personalize Vim in such a way that it becomes your personal favorite. You will find recipes for the following personalization tasks: 1. Changing the fonts 2. Changing the color scheme 3. Personal highlighting 4. A more informative status line 5. Toggle menu and toolbar 6. Adding your own menu and toolbar buttons 7. Work area personalization
Personalizing Vim
Some of these tasks contain more that one recipe because there are different aspects to personalizing Vim for that particular task. It is you, the reader, who decides which recipes (or parts thereof) you would like to use. Before we start working with Vim, there are some things that you need to know about your Vim installation—where to find the configuration files.
Where are the Config Files?
When working with Vim, you need to know a range of different configuration files. The location of these files is very dependent on where you have installed Vim, and the operating system that you are using. In general, there are three configuration files that you must know where to find.
vimrc
This is the main configuration file for Vim. It exists in two versions—global and personal. The global vimrc file is placed in the folder where all your Vim system files are installed. You can find out the location of this folder by opening Vim and executing the following command in normal mode: :echo $VIM
Examples could be: Linux: /usr/share/vim/vimrc Windows: c:\program files\vim\vimrc The personal vimrc file is placed in your home directory. The location of the home directory is dependent on your operating system. Vim originally was meant for UNIXes, so the personal vimrc file is set to be hidden by adding a dot as the first character in the filename. This normally hides files on UNIXes but not on Microsoft Windows. Instead, the vimrc file is prepended with an underscore on these systems. So, examples would be: Linux: /home/kim/.vimrc Windows: c:\documents and settings\kim\_vimrc
[ 16 ]
Chapter 2
Whatever you change in the personal vimrc file will overrule any previous setting made in the global vimrc file. This way you can modify the entire configuration without having to ever have access to the global vimrc file. You can find out what Vim considers as the home directory on your system, by executing the following command in normal mode: :echo $HOME
The vimrc file contains ex (vi predecessor) commands, one on each line, and is the default place to add modifications to the Vim setup. In the rest of the book, this file is just called vimrc. Your vimrc can use other files as an external source for configurations. In the vimrc file, you use the command source like this: source /path/to/external/file
Use this to keep the vimrc file clean, and your settings more structured (more about how to keep your vimrc clean in Appendix B).
gvimrc
The gvimrc file is a configuration file specifically for Gvim. It resembles the vimrc file described above, and is placed in the same locations—as a personal version as well as a global version. For example: Linux: /home/kim/.gvimrc and /usr/share/vim/gvimrc Windows: c:\documents and settings\kim\_gvimrc, and c:\program files\vim\gvimrc This file is used for GUI-specific settings that only Gvim will be able to use. In the rest of the book, this file is called gvimrc
exrc
This is a configuration file that is only used for backwards compatibility with the old vi/ex editor. It is placed at the same location (both global and local) as vimrc and is used the same way. However, it is almost never used anymore except if you want to use Vim in vi-compatible mode.
[ 17 ]
Personalizing Vim
Changing the Fonts
In regular Vim there is not much to do when it comes to changing the font because the font follows the one of the terminal. In Gvim however, you are given the ability to change the font as much as you like. The main command for changing the font in Linux is: :set guifont=Courier\ 14
Where Courier can be exchanged with the name of any font that you have, and 14 with any font size you like (size in points—pt). For changing the font in Windows, use the following command: :set guifont=Courier:14
If you are not sure about whether a particular font is available on the computer or not, you can add another font at the end of the command by adding a comma between the two fonts. For example: :set guifont=Courier\ New\ 12, Arial\ 10
If the font name contains a whitespace or a comma, you will need to escape it with a backslash. For example: :set guifont=Courier\ New\ 12
This command sets the font to Courier New size 12—but only for this session. If you want to have this font every time you edit a file, the same command has to be added to your gvimrc file (without the ':' in front of set). In Gvim on Windows, Linux (using GTK+), Mac OS, or Photon, you can get a font selection window shown if you use the command: :set guifont=*
If you tend to use a lot of different fonts depending on what you are currently working with (code, text, log-files, etc.), you can set up Vim to use the correct font according to the filetype. For example, if you want to set the font to Arial size 12 every time a normal text file (.txt) is opened, this can be achieved by adding the the following line to your vimrc file: autocmd BufEnter *.txt set guifont=Arial\ 12
[ 18 ]
Chapter 2
The window of Gvim will resize itself every time the font is changed. This means that if you use a smaller font you will also (as a default) have a smaller window. You will notice this right away if you add several different filetype commands like the one above, and then open some files of different types. Whenever you switch to a buffer with another filetype, the font will change, and hence the window size too. You can find more information about changing fonts in the Vim help system under: :help 'guifont'
Changing Color Scheme
Often, when working in a console environment you have only a black background and white text in the foreground. This is, however, both dull and dark to look at. Some colors would be desirable. As a default, you have the same colors in console Vim as in the console you opened it from. However, Vim has given its users the opportunity to change the colors it uses. This is mostly done with a color scheme file. These files are usually placed in a directory called colors wherever you have installed Vim. You can easily change among the installed color schemes with the command: :colorscheme mycolors
where mycolors is the name of one of the installed color schemes. If you don't know the names of the installed color schemes, you can place the cursor after writing: :colorscheme
and shift through the names by pressing the tab-key. When you find the color scheme you want, you can press the enter key to apply it. The color scheme does not apply only to foreground and background color, but also to the way code is highlighted, how errors are marked, and other visual markings in the text. You will find that some color schemes are very alike and only minor things have changed. The reason for this is that the color schemes are user supplied. If some user did not like one of the color settings in a scheme, he or she could just change that single setting and re-release the color scheme under a different name.
[ 19 ]
Personalizing Vim
Play around with the different color schemes and find the one you like. Now, test it in the situations where you would normally use it, and see if you still like all the color settings. In Chapter 6, we will get back to how you can change a color scheme to fit your needs perfectly.
Personal Highlighting
In Vim, the feature of highlighting things is called matching.
With matching, you can make Vim mark almost any combination of letters, words, numbers, sentences, and lines. You can even select how it should mark it (errors in red, important words in green, etc). Matching is done with the following command: :match Group /pattern/
The command has two arguments. The first one is the name of the color group that you will use in the highlight. Compared to a color scheme, which affects the entire color setup, a color group is a rather small combination of background (or foreground) colors that you can use for things like matches. When Vim is started, a wide range of color groups are set to default colors, depending on the color scheme you have selected.
To see a complete list of color groups, use the command : :so $VIMRUNTIME/syntax/hitest.vim
[ 20 ]
Chapter 2
The second argument is the actual pattern you want to match. This pattern is a regular expression, and can vary from being very simple to extremely complex, depending on what you want to match. A simple example of the match command in use would be: :match ErrorMsg /^Error/
This command looks for the word Error (marked with a ^) at the beginning of all lines. If a match is found, it will be marked with the colors in the ErrorMsg color group (typically white text on red background).
If you don't like any of the color groups available, you can always define your own. The command to do this is as follows: :highlight MyGroup ctermbg=red guibg=red gctermfg=yellow guifg=yellow term=bold
This command creates a color group called "MyGroup" with a red background and yellow text, in both console (Vim) and GUI (Gvim). You can change the following options according to your preferences: •
ctermbg : Background color in console
•
guibg
•
ctermfg : Text color in console
•
guifg
: Text color in Gvim
•
gui
: Font formatting in Gvim
•
term
: Font formatting in console (for example, bold)
: Background color in Gvim
If you use the name of an existing color group, you will alter that group for the rest of the session. When using the match command, the given pattern will be matched until you perform a new match or execute the following command: :match NONE
[ 21 ]
Personalizing Vim
The match command can only match one pattern at a time; so Vim has provided you with two extra commands to match up to three patterns at a time. The commands are easy to remember because their names resemble that of the match command: :2match :3match
You might wonder what all this matching is good for, as it can often seem quite useless. To show the strength of matching, here are a few examples: Example 1: In mails, it is a common rule that you do not write lines more than 74 characters long (a rule that also applies to some older programming languages like for example Fortran-77). In a case like this, it would be nice if Vim could warn you when you reached this specific number of characters. This can simply be done with the following command: :match ErrorMsg /\%>73v.\+/
Here, every character after the 73rd character will be marked as an error. This match is a regular expression that when broken down consists of: \%> : Match after column with the number right after this 73
: The column number
V
: Combined with the previous command, this means that the next part is very magic. See :help magic for more info.
.\+ : Match one or more of any character.
Example 2: When coding, it is generally a good rule of thumb that tabs are only to be used to indent code, and not anywhere else. However, for some it can be hard to obey this rule. Now, with the help of a simple match command this can easily be prevented. The following command will mark any tabs that are not at the beginning of the line (indentation) as an error: :match errorMsg /[^\t]\zs\t\+/
[ 22 ]
Chapter 2
Now you can check if you have forgotten the rule and used the tab key inside the code. Broken down, the match consists of the following parts: [^
: Begin a group of characters that should not be matched
\t
: The tab-character
]
: End of character group.
\zs : A zero-width match that places the 'matching' at the beginning of the line
ignoring any whitespaces
\t\+ : One or more tabs in a row.
This command says: don't match all the tab-characters, match only the ones that are not used at the beginning of the line (ignoring any whitespaces around it).
If instead of using tabs if you want to use the space character for indentation, then you can change the command to: :match errorMsg /[\t]/
This command just says: match all the tab-characters. Example 3: If you write a lot of IP addresses in your text, sometimes you tend to enter a wrong value in one (like 123.123.123.256). To help you prevent this kind of an error, you can add the following match to your vimrc file: match errorMsg /\(2[5][6-9]\|2[6-9][0-9]\|[3-9][0-9][0-9]\)[.] \[0-9]\{1,3\}[.][0-9]\{1,3\}[.][0-9]\{1,3\}\| \[0-9]\{1,3\}[.]\(2[5][6-9]\|2[6-9][0-9]\|\ \\ \[3-9][0-9][0-9]\)[.][0-9]\{1,3\}[.][0-9] \\{1,3\}\|\[0-9]\{1,3\}[.][0-9]\{1,3\}[.]\(2[5] \\ \[6-9]\|\2[6-9][0-9]|[3-9][0-9][0-9]\)[.] [0-9]\{1,3\} \\|[0-9]\{1,3\}[.][0-9]\{1,3\}[.][0-9]\{1,3\}[.] \\(2[5][6-9]\|2[6-9][0-9]\|\[3-9][0-9][0-9]\)/
Even though this seems a bit too complex for solving a small possible error, you have to remember that even if it helps you just once, it has already been worth adding.
[ 23 ]
Personalizing Vim
If you want to match valid IP addresses, you can use this much simpler command: match todo /\(\(25[0-5]\|2[0-4][0-9]\|[01]\?[0-9] [0-9]\?\)\.\) \\ \{3\}\(25[0-5]\|2[0-4][0-9]\|[01]\? [0-9][0-9]\?\)/
A More Informative Status Line
At the bottom of the Vim editor, you will find two things: the command-line buffer (where you can input commands), and the status line. In the default configuration, Vim has a simple and quite non-informative status line. To the right it shows the number of the current row and column and to the left it shows name of the file currently open (if any). Whenever you execute a Vim command, the status line will disappear and the command buffer will be shown in that line instead. If the command you execute writes any messages, then those will be shown on the right of the status line. For simple and fast file editing, this status line is adequate. But if you use Vim every day and for a lot of different file formats, it would be nice to have a more informative statusline. In this recipe, we see some examples of how the status line can be made a lot more informative with simple methods. The command that sets how the status line should look is simply called: :set statusline format
where format is a printf-like string (known from C programming) that describes how the status line should look. If you look in the Vim help system by typing :help 'statusline', you will see that the status line can contain a wide variety of pieces of information, some more useful in your daily work than others. My status line always contains information about: •
Name of the file that I am editing
•
Format of the file that I am editing (DOS, Unix)
•
Filetype as recognized by Vim for the current file
•
ASCII and hex value of the character under the cursor [ 24 ]
Chapter 2
•
Position in the document as row and column number
•
Length of the file (line count)
The following command will turn your status line into a true information bar with all the above information: :set statusline=%F%m%r%h%w\ [FORMAT=%{&ff}]\ [TYPE=%Y]\ [ASCII=\%03.3b]\ [HEX=\%02.2B]\ [POS=%04l,%04v][%p%%]\ [LEN=%L]
I have added a '[ ]' around each of the pieces of information, so that it is easier to distinguish them from each other. This is purely to give a visual effect and can be left out if necessary.
However, we now see that the status line still shows the old non-informative status line, as in the default installation. This problem occurs because Vim, by default, does not show the status line at all. Instead, it just shows the command buffer with a little bit of information in it. To tell Vim that you would like to have a real status line shown, you will have to add the following setting to your vimrc file. This command will make sure that your status line is always shown as the second last line in the editor window: :set laststatus=2
You will now see that the command buffer gets a place of its own in the last line of the editor window. This way there's always room for the status line and you will always have information about the file right in front of you. The status line does of course take up some of the editing area, but it is then up to you to decide whether it should be shown or not. You can always remove it for the rest of the editing session by executing the following command from within Vim: :set laststatus=0
Toggle Menu and Toolbar
If you are used to working with Vim in the console mode, you are also quite used to having no menus and toolbars in the top of the window. However, when you move to Gvim, you will soon realize that both the menu and the toolbar are there, by default, in the GUI.
[ 25 ]
Personalizing Vim
Many believe that extra room for text is far more important than the menu and the toolbar. If you are one of those persons, you might like to remove the menu and toolbar while working in Gvim. However, some scripts add useful functionality in the menu and it is therefore important to have the menus. The solution for this could be toggling if the menu and toolbar is shown or not. The following code maps the key combination Ctrl-F2 to toggle the menu and toolbar in Gvim. You can add it to your vimrc file if you want this functionality. map <silent> :if &guioptions =~# 'T' \set ������������������������ guioptions-=T \set guioptions-=m \else \set ������������������������ guioptions+=T \set guioptions+=m \endif
Now, whenever you don't need the menu and toolbar, you can just press Ctrl-F2 and you will get the full space for your text. If you want either the menu or the toolbar to be hidden all the time, add one of the following lines to your vimrc file:. To remove the menu completely: :set guioptions-=m
To remove the toolbar completely: :set guioptions-=T
Other parts of the GUI can be modified with the set guioptions command. To find out what you can modify, look in :help 'guioptions'
Adding Your Own Menu and Toolbar Buttons
If you are in Gvim, you can make a handy menu with all the functionality you use the most. You might not always need to use it from the menu, but whenever you forget how to use it, you can always just find it there. If you need to get to the functionality really fast, you can even add it directly in the toolbar of Gvim.
[ 26 ]
Chapter 2
In this recipe, we look at both how to make your own menu and, later, how to add extra buttons to the toolbar in Gvim. Let us start with the menu construction.
Adding a Menu
Building a menu is basically just executing a command for each item you want in the menu. As long as you follow the right naming convention, you will see a nice little menu with all your items in it. Let us start with a simple example. Say you want to add a menu like the buffers menu, but for tabs.
The command you will need to use is: :menu menupath command
This command works much like the map command, except that instead of mapping a command to a key combination, here the mapping is done to a menu item. The command has two arguments. The first is the actual path in the menu where the item should be placed, and the second argument is the command that the menu item should execute. If for instance, you want to add a menu item called Next to the menu item Tabs, then you would need to use a command like this: :menu Tabs.Next <ESC>:tabnext
So now you have a menu called Tabs with one menu item called Next. What the Next menu item does is execute the following command: :tabnext
[ 27 ]
Personalizing Vim
This command is prepended with <Esc> to get into the normal mode, and then to actually execute the command. If you haven't added <Esc> this command won't work. Another way to get around this is by adding specific menu items according to the current mode. For this Vim has a range of alternatives to the :menu command: :nmenu – for Normal mode. :imenu – for Insert mode. ^O is prepended. :vmenu – for Visual mode. ^C is prepended and ^\^G is appended. :cmenu – for Command-line mode. ^C is prepended and ^\^G is appended. :omenu – for OP-pending mode. ^C is prepended and ^\^G is appended.
The prepended parts (^O and ^C) are to get into normal mode. The ^O (Ctrl-O) is especially for insert mode because it gets you back into insert mode after executing the command. ^\^G (Ctrl-\, Ctrl-G) is to handle the special case wherein the global insert
mode setting is set to true and Vim has insert mode as the default mode (Vim is mode-less). In this case, it will get you back into insert mode and in the rest of the cases it will get you back in the mode you just came from. Instead of setting the same menu item for each and every mode, you can just replace the commands with this single command: :amenu menu-path command
According to the current mode, this command prepends and appends the right things.
So let's go to our new Tabs menu, and add some more items and functionality to it. With the following, it should look similar to the Buffers menu: :amenu :amenu :amenu :amenu
Tabs.&Delete :confirm tabclose Tabs.&Alternate :confirm tabn # <silent> Tabs.&Next :tabnext <silent>Tabs.&Previous :tabprevious
The observant reader might have noticed that some new things have been added in the commands. The first thing is the <silent> tag in the last two commands. By adding this we can avoid the command being echoed in the command-line buffer during execution. While this is a purely cosmetic functionality, the '&' in the menu path is a more functional extension. By adding an '&' in front of one of the letters in the last part of the menu path, you can define a keyboard shortcut for an item. This makes it easy to navigate to that particular item in the menu and execute it. [ 28 ]
Chapter 2
Let's say that you want to go to the next tab by executing the Tabs > Next menu item; now you can do so by simply pressing Alt-t n. This is Alt-t for opening Tabs, and n to call the Next item—n because the '&' is in front of the N in Next. If another menu item uses the same character for a shortcut, you can cycle through them by pressing the Alt key repeatedly. If you would like to have a line that separates some of the items in your menu drop down, you can use the name 'SEP' for the item and ':' for the command: :amenu Tabs.-SEP:
The menu that we have created will only exist as long as Vim is open in this session, so in order to get it into your menu all the time, you need to add it to your vimrc file (without the ':' in front of the commands).
So now we have a simple tabs menu that looks a bit like the Buffers menu. It does not, however, have the functionality that lists active buffers in the Buffers menu. This does not make much of a difference when you realize that buffers can be hidden for the user, but tabs cannot. You can, in other words, always see exactly how many tabs you have and what they are called by just looking at the tab bar. A personal menu can be used for a lot of other interesting things. If you work with many types of files you can even start having menus for specific file types or sub-menus for the different types in the same menu. A sub-menu is constructed by following the naming convention in the menu path. So if you want to have Tabs > Navigation > Next, you will simply have to add the Next menu item with the menu path Tabs.Navigation.&Next
[ 29 ]
Personalizing Vim
Adding Toolbar Icons
So now that we know how to make our menus, adding our own icons to the toolbar isn't that difficult. Actually, Vim is constructed in such a way that the toolbar is just another menu with a special name. Hence, adding an icon to the toolbar is just like adding an item to a menu. In the case of a 'toolbar menu', you will be able to add items to it by using a menu-path that starts with the name ToolBar. To add an item to the toolbar that gives access for executing the command :buffers (show list of open buffers), all you have to do is to execute the following command: :amenu icon=/path/to/icon/myicon.png ToolBar.Bufferlist :buffers
Of course, you will need to have an icon placed somewhere that can be shown in the toolbar. The path to the icon is given with the argument icon to the amenu command. If you do not give a path to the file, but only the filename, then Vim will look for the icon in a folder called bitmaps/ in the Vim runtimepath (execute :echo $VIMRUNTIME to see where it is). The type of icons supported is dependant on the system you use it on. And that's really it! After executing the command, you will see your icon in the toolbar as the last one on the right. If you press it, it will execute the command, :buffers, and show you a buffer list. As with the menus, you can add toolbar buttons that are only shown in specific modes using the mode-specific menu commands imenu, vmenu, cmenu, etc. If you want your menu or toolbar icon placed elsewhere than to the right of the others, then you can use priorities. Read more about how in: :help menu-priority and :help sub-menu-priority
Modifying Tabs
Ever since the release of Vim version 7.0, there has been support for tabs or tab pages as it is called. Tab pages are not like the normal tabs in other applications; rather they are a way to group your open files. Each tab can contain several open buffers and even several windows at the same time. What makes tabs special is that the commands you would normally execute on all open buffers/windows (like :bufdo,:windo, :all,:ball) are limited to only the windows and buffers in the current tab page. [ 30 ]
Chapter 2
Normally, tab pages are shown as a list of tabs in the top of the window (just above the editing area). Each tab has a label, which as a default shows the name of the file in the currently active buffer. If more windows are open, at the same time, in the tab page, then the tab label will also show a number telling how many windows.
Sometimes you might like to have the label on the tabs telling you something different. For instance, if you often have one tab for each project, then it would be nice to name the tab according to the name of the project in it. The label on the tabs is set in a way very much similar to the one used for the status line (see section A More Informative Status Line). But here, instead of setting the property status line, you set the property tabline: :set tabline=tabline-layout
or if you are in Gvim: :set guitablabel
Even though setting the tabline resembles the way you set the status line, it is a bit more troublesome. This is mainly because you need to take care of whether the tab is the active one or not. So let's start with a little example for Vim. When we have a lot of tabs, they tend to take up too much space in the tab page, especially if they contain the entire name of the file in the currently active buffer. We want to have only the first 6 letters of the name of the active buffer in the tab label. The active tab should also be easy to distinguish from the other tabs; so let's make its colors white on red like error messages. The following script in Vim script does just that (learn more about how to create Vim scripts in Chapter 6). function ShortTabLine() let ret = '' for i in range(tabpagenr('$')) " select the color group for highlighting active tab if ����������������������� i + 1 == tabpagenr() let ret .= '%#errorMsg#' else
[ 31 ]
Personalizing Vim let ret .= '%#TabLine#' ������������ endif " find the buffername for the tablabel let buflist = tabpagebuflist(i+1) let winnr = tabpagewinnr(i+1) let buffername = bufname(buflist[winnr – 1]) let filename = fnamemodify(buffername,':t') " check if there is no name if filename == '' let filename = 'noname' endif " only show the first 6 letters of the name and " .. if the filename is more than 8 letters long if ����������������������� strlen(filename) >=8 let ret .= '['. filename[0:5].'..]' else let ret .= '['.filename.']' ���������������� endif endfor " after the last tab fill with TabLineFill and reset tab page # let ret .= '%#TabLineFill#%T' return ret endfunction
Now, we have the function and just need to add it to our vimrc file, along with a line that sets the tabline to the output of our function. This can be done with the following command: :set tabline=%!ShortTabLine()
The result is a more compact tablist as shown in the following screenshot:
Changing the tabline in Gvim is a bit different, but still follows almost the same basic ideas. However, when in the GUI, you do not have to consider things like the color of the active tab, or whether it is actually active or not because this is all a part of the GUI design itself. So let's simplify the ShortTabLine() function a bit so that it only sets the tab label: function ShortTabLabel() let bufnrlist = tabpagebuflist(v:lnum) [ 32 ]
Chapter 2 " show only the first 6 letters of the name + .. let label = bufname(bufnrlist[tabpagewinnr(v:lnum) – 1]) let filename = fnamemodify(label,':h') " only add .. if string is more than 8 letters if strlen(filename) >=8 let ret=filename[0:5].'..' else let ret = filename endif return ret endfunction
So now we just have to set the guitablabel property to the output of our function: :set guitablabel=%{ShortTabLabel()}
The result will be fine small tabs as shown in the following figure.
If you want to remove the tabs bar completely from Gvim, then you can use the command::set showtabline=0 (set to 1 to get it shown again).
So now we have limited the information in the tabs, but we would still like to have the information somewhere. For that we have a nice little tip—use the tool tips. The nice thing about tool tips is that when you don't activate them (hold your cursor over some area, e.g., a tab) you don't see them. This way you can have the information without it filling up the entire editor. To set the tool tip for a tab you will need to use the following command: :set guitabtooltip
This property should be set to the value you want to show, when the mouse cursor hovers over the tab. To test it you can try with a simple execution like: :set guitabtooltip='my tooltip'
[ 33 ]
Personalizing Vim
Now, this only shows a static text in the tool tip. We need some more information there. We removed the path from the filenames on the tabs, but sometimes it is actually nice to have this information available. With the tool tips this is easily shown with the following command: :set guitabtooltip=%!bufname($)
As with the tabs, the contents of the tool tip can be constructed by a function. Here we have constructed a small function that shows all the information you would normally have in the tabs—but in a more organized way: function! InfoGuiTooltip() "get window count let wincount = tabpagewinnr(tabpagenr(),'$') let bufferlist='' "get name of active buffers in windows for i in tabpagebuflist() let bufferlist .= '['.fnamemodify(bufname(i),':t').'] ' endfor return bufname($).' windows: '.wincount.' ' .bufferlist ' ' endfunction
You use this code described above like this: :set guitabtooltip=%!InfoGuiTooltip()
You can probably imagine many other interesting ways to use the small information space the tabs, and tool tips provide, and following the above example, you should have no problems in implementing them.
Work Area Personalization
In this section, we introduce a list of smaller, good-to-know, modifications for the editor area in Vim. The idea with these recipes is that they all give you some sort of help or optimization when you use Vim for editing text or code. [ 34 ]
Chapter 2
Adding a More Visual Cursor
Sometimes, you have a lot of syntax coloring in the file you are editing. This can make the task of tracking the cursor really hard. If you could just mark the line the cursor is currently in, then it would be easier to track it. Many have tried to fix this with Vim scripts but the results have been near useless (mainly due to slowness, which prevented scrolling longer texts at an acceptable speed). Not until version 7 did Vim have a solution for this, but then it came up with not just one, but two possible solutions for cursor tracking. The first one is the cursorline command, which basically marks the entire line with, for example, another background color, without breaking the syntax coloring. To turn it on, use the following command: :set cursorline
The color it uses is the one defined in the color group CursorLine. You can change this to any color or styling you like, for example: :highlight CursorLine guibg=lightblue ctermbg=lightgray
See the section Personal Highlighting for more info on how to change a color group.
If you are working with a lot of aligned file content (like tab-separated data), the next solution for cursor tracking comes in handy: :set cursorcolumn
This command marks the current column the cursor is in by, for example, coloring the entire column in the file. As with the cursorline, you can change the settings for how the cursor column should be marked. The color group to change is named CursorColumn.
[ 35 ]
Personalizing Vim
Adding both the cursor line and column marking makes the cursor look like a crosshair, thus making it impossible to miss.
Warning! Even though the cursorline and cursorcolumn functionality is implemented natively in Vim, it can still give quite a slowdown when scrolling through the file.
Adding Line Numbers
Often when compiling and debugging code, you will get error messages stating that the error is in some line. One could of course start counting lines from the top to find the line, but Vim has a solution to go directly to some line number. Just execute :XXX where XXX is the line number, and you will be taken to line XXX. Alternatively, you can go into normal mode (press Esc) and then simply use XXXgg or XXXG (again XXX is the line number). Sometimes, however, it is nice to have an indication of the line number right there in the editor, and that's where the following command comes in handy: :set number
Now you get line numbers to the left of each line in the file. By default, the numbers take up four columns of space—three for numbers and one for spacing. This means that the width of the numbers will be the same until you have more than 999 lines. If you get above this number of lines, an extra column will be added and the content will be moved to the right. You can of course change the default number of columns used for the line numbers. This can be achieved by changing the following property: :set numberwidth=XXX
Replace XXX with the number of columns that you want.
[ 36 ]
Chapter 2
Even though it would be nice to make the number of columns higher in order to get more spacing between code and line numbers, this is not achievable with the numberwidth property. This is because the line numbers will be right aligned within the columns.
You can change the styling of the line numbers, and the columns they are in, by making changes to the color group LineNr.
Spell Checking Your Language
We all know it! Even if we are really good spellers, it still happens from time to time that we misspell a word or hit the wrong keys. In the past, you had to run your texts (that you had written in Vim) through some sort of spell checker like Aspell or Ispell, which was a tiresome process that could only be performed as a final task—unless you wanted to do it over and over again. With version 7 of Vim, this troublesome way of spell checking is over. Now, Vim has got a built-in spell checker with support for more than 50 languages from around the world. The new spell checker marks the wrongly written words as you type them in, so you know there is an error right away. The command to execute to turn on this helpful spell checker feature is: :set spell
This turns on the spell checker with the default language (English). If you don't use English much, and would prefer to use another language in the spell checker, then there is no problem changing this. Just add the code of the language you would like to use to the spelllang property, for example: :set spelllang=de
[ 37 ]
Personalizing Vim
Here, the language is set to German (Deutsch) as the spell checker language of choice. The language name can be written in several different formats. American English, for example, can be written as: •
en_us
•
us
•
American
Names can even be an industry-related name like 'medical'. If Vim does not recognize the language name, it will be highlighted when you execute the property-setting command. If you change the spelllang setting to a language not already installed, then Vim will ask you if it should try to retrieve it from the Vim homepage, automatically.
Personally, I tend to work in several different languages in Vim, and I really don't want to tell Vim all the time which language I am using right now. Vim has a solution for this. By appending more language codes to the spelllang property (separated by commas), you can tell Vim to check the spelling in more than one language. :set spelllang=en,da,de,it
Vim will then take the languages from the start to the end, and check if the words match any word in one of these languages. If they do, then they are not marked as a spelling error. Of course, this means that you can have a word spelled wrong in the language you are using but spelled correctly in another language, thereby introducing a hidden spelling error. You can find language packages for a lot of languages at the Vim FTP site: ftp://ftp.vim.org/pub/vim/runtime/spell
[ 38 ]
Chapter 2
Spelling errors are marked differently in Vim and Gvim. In regular Vim, the misspelled word is marked with the SpellBad color group (normally white on red). In Gvim, the misspelled word is marked with a red curvy line underneath the word. This can of course be changed by changing the settings of the color group (See the section Personal Highlighting for more info).
Whenever you encounter a misspelled word, you can ask Vim to suggest better ways to spell the word. This is simply done by placing the cursor over the word and then going into the normal mode (press Esc), and then pressing z=.
Vim, if possible, will give you a list of good guesses for the word you were actually trying to write. In front of each suggestion is a number. Press the number you find in front of the right spelling (of the word you wanted) or Enter if the word is not there. Often Vim gives you a long list of alternatives for your misspelled word, but unless you have spelled the word completely wrong, chances are that the correct word is within the top 5 of the alternatives. If this is the case, and you don't want to look through the entire list of alternatives, then you can limit the output with the following command: :set spellsuggest=X
Set X to the number of alternative ways of spelling you want Vim to suggest.
[ 39 ]
Personalizing Vim
Adding Helpful Tool Tips
In the recipe Modifying Tabs, we learned about how to use tool tips to store more information in the tabs in Gvim, without taking up much space. To build on top of that same idea, with this recipe, we move on and use tool tips in other places in the editor. The editing area is the largest part of Vim; why not try to add some extra information to the contents of this area by using tool tips? In Vim, tool tips for the editing area are called balloons and they are only shown when the cursor is hovering over one of the characters. The commands you will need to know in order to use the balloons are: :set ballooneval :set balloondelay=400 :set ballonexpr="textstring"
The first command is the one you will use to actually turn on this functionality in Vim. The second command tells Vim how long it should wait before showing the tool tip/ balloon (the delay is in milliseconds and as a default is set to 600). The last command is the one that actually sets the string that Vim will show in the balloon. This can either be a static text string or the return of some function. In order to have access to information about the place where you are hovering over a character in the editor, Vim gives access to a list of variables holding such information: •
v:beval_bufnr : Number of the buffer in which the hovered area is.
•
v:beval_winnr : Number of the window in which the hovered area
•
v:beval_lnum
: Line number on which the hovered character is situated.
•
v:beval_col
: Number of the column in which the hovered character is.
•
v:beval_text
: Word to which the hovered character is connected.
is shown.
So with these variables in hand, let's look at some examples. Example 1: The first example is based on one from the Vim help system, and shows how to make a simple function that will show the info from all the available variables. function! SimpleBalloon() return 'Cursor is at line/column: ' . v:beval_lnum . [ 40 ]
Chapter 2 \'/' . v:beval_col . \ ' in file ' . bufname(v:beval_bufnr) . \ '. Word under cursor is: "' . v:beval_text . '"' endfunction set balloonexpr=SimpleBalloon() set ballooneval
The result will look like in the following figure:
Example 2: Let's look at a more advanced example that explores the use of balloons for specific areas in editing. In this example, we will put together a function that gives us great information balloons for two areas at the same time: •
Misspelled words—the balloon gives ideas for alternative words.
•
Folded text—the balloon gives a preview of what's in the fold.
So let's take a look at what the function should look for, to detect if the cursor is hovering over either a misspelled word, or a fold line (a single line representing multiple lines folded behind it). In order to detect if a word is misspelled, the spell check would need to be turned on: :set spell
If it is on, then calling the built-in spell checker function, spellsuggest(), would return alternative words if the hovered word was misspelled. So, to see if a word is misspelled is just to check if the spellsuggest() returns anything. There is, however, a small catch. spellsuggest() also returns alternative, similar words, if the word is not misspelled. To get around this, another function has to be used on the input word before putting it into the spellsuggest() function. This extra function is the spellbadword(). This basically moves the cursor to the first misspelled word in the sentence that it gets as input, and then returns the word. We just input a single word, and if it is not misspelled, then the function cannot return any words. Putting no word into spellsuggest() results in getting nothing back, so we can now check if a word is misspelled or not. [ 41 ]
Personalizing Vim
To check if a word is in a line, in a fold, is even simpler. You simply have to call the function foldclosed() on the line number of the line over which the cursor is hovering (remember v:beval_lnum ?) and it will return the number of the first line in the current fold—if not in a fold, then it returns -1. In other words, if foldclosed(v:beval_lnum) returns anything but -1 and 0, we are in a fold. Putting all of this detection together and adding functionality to construct the balloon text ends up as the following function: function! FoldSpellBalloon() let foldStart = foldclosed(v:beval_lnum ) let foldEnd = foldclosedend(v:beval_lnum) let �������������� lines = [] " Detect if we are in a fold if foldStart < 0 " Detect if we are on a misspelled word let lines = spellsuggest( spellbadword(v:beval_text)[ 0 ], 5, 0 ) else " we are in a fold let numLines = foldEnd - foldStart + 1 " if we have too many lines in fold, show only the first 14 " and the last 14 lines if ( numLines > 31 ) let ������������������������������������������������ lines = getline( foldStart, foldStart + 14 ) let ���������� lines += ������������������������������������������������������ [ '-- Snipped ' . ( numLines - 30 ) . ' lines --' ] let ���������� lines += ����������������������������������� getline( foldEnd - 14, foldEnd ) else ���� "less than 30 lines, lets show all of them let ����������������������������������������� lines = getline( foldStart, foldEnd ) endif ����� endif " return result return join( lines, has( "balloon_multiline" ) ? "\n" : " " ) endfunction set balloonexpr=FoldSpellBalloon() set ballooneval
[ 42 ]
Chapter 2
The result is some really helpful balloons in the editing area of Vim that can improve your work-cycle tremendously. The following figure shows how the info balloon could look when using it to preview a folded range of lines from a file.
If the balloon is instead used on a misspelled word, it will look like this:
In Chapter 4, you can learn more about how to use folding of lines to boost productivity in Vim.
Using Abbreviations
We all know the feeling of writing the same things over and over again, a dozen times during a day. This feeling is the exact opposite of that the philosophy of Vim tries to teach us. The philosophy of Vim says that if you write a thing once, it is OK, but if you're writing it twice or more times, then you should find a better way to do it.
[ 43 ]
Personalizing Vim
One of the methods for getting around writing the same sentences over and over again is by using abbreviations. In Vim, abbreviations are created with one of the following commands depending on which mode they should be available in: :abbreviate
: Abbreviations for all modes
:iabbrev
: Abbreviations for insert mode
:cabbrev
: Abbreviations for the command line only
All of the commands take two arguments: the abbreviation, and the full text it should expand to. So let's start with a simple example of where the abbreviations can come in handy. Example 1: I have moved around a bit during the last few years, so a common task for me is writing messages where I tell about my new address. It didn't take me long before I had an abbreviation ready, so I didn't have to write the entire address. Here is what it looked like: :iabbrev myAddr 32 Lincoln Road, Birmingham B27 6PA, United Kingdom
So now, every time I need to write my address, I just write myAddr, and as soon as I press space after the word, it expands to the entire address. Vim is intelligent about detecting whether you are writing an abbreviation or it is just part of another word. This is why myAddr only expanded to the full address after I pressed space after the word. If the character right after my abbreviation was a normal alphabetic letter, then Vim would know that I wasn't trying to use my abbreviation and it would not expand the word. Examples with the abbreviation 'abc': • • • •
abc<space> and abc<enter> : Both expand. 123abc<space> : Will not expand since abbreviation is part of a word. abcd<space> : Will not expand because there are letters after the abbreviation. abc : Will not expand until another special letter is pressed.
[ 44 ]
Chapter 2
A good place to keep your abbreviations, so that you don't have to execute all the commands by hand is in a file in your VIMHOME. Simply place a file there (let's call it abbreviations.vim) and write all your abbreviations in it. Then, in your vimrc file, you just make sure that the file is read, which is done with the source command: :source $VIM/abbreviations.vim
Every time you realize that you will need a new abbreviation, you first execute it, and then you add it to your abbreviations.vim. By now you have probably realized that you can use abbreviations for a lot of other interesting things. But anyway here is a short list of examples to give you some ideas: •
Correct typical keypress errors: :iabbr teh the
•
Simple templates for programming: :iabbr forx
•
for(x=0;x:wa
So what happens now, is that you map the Ctrl-s to do a combination of key presses. First, <esc> (the Escape key), to get out of insert mode and into normal mode. Then, :w to execute the actual saving of the file, and finally the a, to get back into insert mode and go to the end of the line. You could expand the mappings to fit all of the standard copy/paste/cut/save shortcuts from many applications. This could be constructed like: " save file (ctrl-s) :map :w " copy selected text (ctrl-c) :vmap y " Paste clipboard contents (ctrl-v) :imap <esc>P [ 46 ]
Chapter 2 " cut selected text (ctrl-x) :vmap x
If you are in Gvim, you can even get dialogs shown for the Save-as and Open functionalities. "Open new file dialog (ctrl-n) :map :browse confirm e "Open save-as dialog (ctrl-shift-s) :map :browse confirm saveas
With the ability to change the keyboard mapping in Vim, you really have access to a powerful way of modifying the editor completely according to your needs. You can read more about mappings in the vim help system under: :help key-mapping
Summary
In this chapter, we have looked at how to make Vim a better editor for you by modifying it to your personal needs. We started out by learning about how basic modifications of font and color scheme can give you editor a personalized look. Then we dived a bit deeper into using colors for marking search matches, thereby making them easily recognizable. To get the most out of an editor like Vim, you would often like it to have a large area for editing the files, and less space spilled on GUI. We looked at ways of modifying both the status line and tabs to be smaller and more informative. If you don't want the menu and toolbar at all, you have also been shown a way for toggling its visibility. Even though the menu and toolbar can be in the way, they can also be very usable additions to your editor. In this chapter, we have learned how to add our own menu to the menu bar and even how to add icons full of functionality to the toolbar. Many things can be done to the editing area to make it fit your personal needs. In this chapter we have looked at how to make it easier to get an overview of the editing area. Better and more visual cursors have been proposed and line numbers have been added to the area.
[ 47 ]
Personalizing Vim
If you need help with your spelling, then Vim has methods for helping you there. We have looked at how to make the spell checker in Vim follow your preferred language, so that you will never again misspell a word. If using spell-checking is not enough to correct your errors, then maybe the use of abbreviations can help you. On the other hand, abbreviations also do a great job minimizing the number of characters to write if you use the same text over and over again. Finally, we have looked at how we can change the key bindings in Vim in such a way that it will react on keyboard shortcuts you are used to from other editors. With all the recipes in this chapter, you should have a fully personalized Vim editor, and you are now ready to move on and learn more about how you can optimize your navigation around the files in Vim.
[ 48 ]
Better Navigation Working with large files, or many files at the same time, can be a troublesome task. Sometimes, you realize that you waste more time looking for content to edit than doing the actual editing. The philosophy of Vim is all about not wasting our valuable time, so Vim has means for optimizing the way we navigate files. In this chapter, we will look at some of the ways in which Vim helps us easily navigate through our files, whether we're dealing with one file or fifty. Some recipes use marks to mark a spot for later return, while other uses search techniques to find the place you are looking for. The recipes in this chapter cover the following areas: 1. Faster navigation in a file 2. Faster navigation in the Vim help system 3. Faster navigation in multiple buffers 4. Faster lookup of files using the Vim file explorer 5. In-file searching 6. Searching in multiple files or buffers with vimgrep 7. Using marks as a tool for navigation 8. Using signs as a tool to get better overview After reading this chapter, you should be able to boost your navigation speed, and have no problems finding the files you are looking for.
Better Navigation
Faster Navigation in a File
Sometimes even the simplest of tasks like navigating through a single file can be optimized. Vim offers several methods of navigation within a file, which can adapt to the contents of the file and how it is organized. Some of these methods are obvious, while others are more complex.
Context-Aware Navigation
Mostly, the files we are editing are well structured. If our files are text, then this structure can be in the form of paragraphs, sentences, and words, or at other times code with functions, blocks, and code lines. Vim supports jumping around the file, according to the structure in the file, and has key bindings that make it easy to go to the exact place in the file where you want to go. Let's look at some examples: Example 1—Moving within a text file: You are working on a normal text file and in the middle of a sentence you realize that you have forgotten to make the first letter in the paragraphs uppercase. You could of course, use the arrow-keys or the h/j/k/l navigation keys to move to the beginning of the paragraph to correct this. It would, however, be faster to just press the following in normal mode: {
You are now placed at the beginning of the paragraph—or in the empty line just above it (if one exists). Now, you go into normal mode by pressing the Esc key and then use the movement command { to go to the beginning of the paragraph. The same can be applied if you want to go back to the end of the paragraph—you simply have to use the counterpart of {, namely: }
Maybe you were not actually working at the end of the paragraph, but rather on correcting some text in the middle of the paragraph. Vim remembers where you were making changes to the file previously (and actually up to 999 of the last places you have changed something) and you can just ask it to take you back to the correct place. Just use the following command in normal mode: g,
[ 50 ]
Chapter 3
Pressing this command several times in a row will loop you through locations of previous changes in the file. As with the { command, this command also has a counterpart that moves forward through the list of recent locations where changes have taken place. The command for this is: g;
Vim will alert you if you get to one of the ends of the list of changes. Maybe it wasn't at the beginning of the paragraph that you had forgotten to capitalize a letter, but rather at the beginning of the current sentence. Again, Vim helps you move faster and offers you a pair of commands to move to the beginning and end of the current sentence. The commands are as follows: (
: Move to the beginning of the sentence.
)
: Move to the end of the sentence.
Vim doesn't want us to waste any time when working in it. Even though you could easily go through the letters of a word by simply using the arrow keys, Vim still thinks this is waste of key-presses and instead offers a set of commands for word movement: w
: Move to the beginning of the next word.
b
: Move to the beginning of the previous word.
e
: Move to the end of the word.
These commands can be combined such that if you want to go to the end of the next word you simply press: we
When it comes to what a word actually consists of, Vim has two definitions. In Vim we have: •
A word consisting of alphabetic letters, numbers, dashes, and underscores
•
A WORD consisting of any character except white spaces (tab and space)
The above-mentioned movement commands work on word, but of course Vim has the same commands available for WORD. Simply use the same commands, but use them in uppercase instead (e.g., W to go to the beginning of the next WORD).
[ 51 ]
Better Navigation
If you want to execute one of the commands mentioned in this section more than once in a row, simply add the number of times you want it executed in front of the command. For example, 5g, to go to the place you changed something 5 changes ago. Example 2 – Moving in a Code file:
Compared to text files, code does not have any paragraphs or sentences to navigate through. It does, however, frequently contain a lot of structures and blocks, each of which has a very specific contextual meaning within the code. An example could be the simple code block: If( a == b) { print "a and b are the same"; }
Here the line with print is within the context of the if block surrounding it. Because Vim is the favorite editor of many programmers, it offers a lot of movement commands to use when you are working with code. Common for all of them is that the parts of the code you want to jump between need to have a contextual connection to each other. A simple example could be a construction like the #if-#else-#endif construction from the C programming language. Here we have a beginning (#if), an end (#endif) and a midpoint (#else). So if you are in the line with #if and press the following command: %
you will go to the #else part. Pressing it again, will take you to the #endif, and yet another execution of the command will get you back to the #if. Vim does not know all programming language constructs, but by default it knows most of the contextual constructs of the C programming language. Besides this, it knows the normal block construction methods from most programming languages—the use of parentheses and brackets (for example, '{' is block start and '}' is block end). If you want Vim to know the constructs of many other programming languages, then install the MatchIt plugin. This plugin is available with Vim as of version 7.0, but can also be found on: http://www.vim.org/scripts/ [ 52 ]
Chapter 3
Simply by knowing the programmer's common use of parentheses/brackets, Vim can provide us with several useful navigation commands. This means that as long as the code uses some start parenthesis/bracket to begin a block, and the counterpart to end it, Vim will understand it. Let's say you are in a function that consists of many lines and you want to go to the beginning of the function. Mostly the brackets surrounding the contents of a function are the outermost pair of brackets around where you currently are in the file (given that you are editing the current function). So for Vim to find the beginning of the function, it simply has to find the outermost bracket pair and then go to the opening bracket. function myExample() { ...many lines of code... /* cursor is placed at the beginning of this line */ ...many lines of code... }
In the above example, the % command would take us to the closing bracket and pressing it again would take us to the opening bracket. But what if the cursor was actually placed inside another pair of brackets? In that case, the % command would only move the cursor to them, and not the beginning of the function. Again Vim has some handy commands for you: [[ and ][
: Move backwards/forward to the next section beginning (e.g., start of a function)
[] and ]]
: Move backwards/forward to the next section end (e.g., end of a function)
Executing these commands multiple times in a row takes you to the beginning/end of the next/previous section, and therefore gives you a convenient way of cycling through the functions in a file. Note that in most object-oriented languages, the Class beginning/end is often the outermost section.
Often, you just want to go to the beginning of the current block (e.g., the beginning of a while-loop) because it is here that you have defined all the local variables for the scope of this block. For this also, Vim has a set of movement commands: [{
: Move to the beginning of the block
]}
: Move to the end of the block [ 53 ]
Better Navigation
If the block in the code is a comment, then it does not have any brackets around it; hence Vim cannot use the brackets to navigate to its beginning/end. Therefore Vim has some special movement commands for comment blocks: [/
: Move to the beginning of the comment block
]/
: Move to the end of the comment block
By default, not all comment formats are supported by Vim. It supports the comment formats used in the C programming language (/* */), C++ (//) and in many scripting languages (#). It is, however, possible to add support for extra comment formats when you add support for the syntax of new programming languages. Sometimes, when you work on a piece of code, you tend to forget how a variable is actually defined. Vim has a command that can help you look up the definition of the variable (or the first occurrence of it, in the case of interpreted languages like Python) if it is defined in the current file. The command is as follows, and should be pressed while the cursor is placed on the variable name you want to look up: gd
This command is easy to remember if you just think of the phrase 'Goto Declaration' and take the first letters from the two words. What this command actually does is start by going to the beginning of the current section (remember the [[ command) because this is where the local definitions are normally placed. Then, it makes a search forward in the file for the first occurrence of the variable name. If it does not find it before reaching the place where you started the lookup, then it moves to the first line of the file and again searches forward in the file looking for any global definitions of the variable. If it still does not find the definition, then Vim does a * search for the variable in the file (read more about the * search in the section Search and You Will Find). If you know that the variable is globally defined, or if you want the global definition of the variable, then Vim has a command that starts by looking from line 1 of the file, instead of looking in the current section first. The command is: gD
Vim is naturally smart enough to ignore any references to the variable in comment blocks because these are definitely not the declaration of the variable. If Vim finds the variable definition (or the first available usage of the variable in the file), then it moves the cursor to this place.
[ 54 ]
Chapter 3
Put a 1 in front of the gd command (like 1gd) if you want Vim to ignore all matches that are inside a { } block that ends before the current cursor position (for example, in another function block earlier in the file).
Navigating Long Lines
Some like their long lines visually wrapped in Vim while others want them to extend beyond the border of the editor (i.e., not shown). Personally, I like to have my text lines wrapped because it makes the overview of the text a lot better. This does, however, introduce an irritating problem. If you have a long line and it is wrapped, then the wrapped part shows up as a new line in the visualization of the file contents. This is not a problem in itself, but navigating to the wrapped part of the line is. If you use the arrow keys, or j/k to navigate between the lines, then it simply ignores the wrapped part of the line and goes directly to the next actual line in the file. If you dislike this behavior, then here's a short little recipe to fix this problem. I have chosen that if I hold down the Alt key, while I use the up/down arrows to navigate the lines in the file, then Vim should follow the lines as shown visually in Vim, and not as the actual lines in the file. The key mappings to make this work are as follows and should simply be added to your vimrc file: map gj map gk imap <ESC>gki imap <ESC>gji
The mapping works in the normal mode and insert mode. If you want this to happen without having to remember to hold down the Alt key, then simply remove the A- part of the key combination to which the commands are mapped (e.g., map gj).
Faster Navigation in Vim Help
Vim comes with a very useful and comprehensive help system that you have probably by now already played around with. What you might not know, however, is that the help system comes with hyperlink support that resembles the hyperlinks we know from the Internet. There are two types of links—subject links marked as 'some subject' and option links marked as 'option'. A subject link refers to the beginning of a section in the help system, whereas the option links takes you directly to the description of a certain option. When you place the cursor on a link, you can press Ctrl-] to follow the link no matter what type of link it is. This is very nice, but if you are using a non-English keyboard layout, the key for ] is often not [ 55 ]
Better Navigation
available with a single key press. In this case, it could be nice to remap the key to some other key. In an Internet browser, you could navigate to a link and press Enter, and to reflect this you could have a mapping like the following: nmap
If you are in a browser and want to return to the previous page you visited, then you can press the Backspace key. This feature would be nice to have in the Vim help system; hence a mapping like the following could be useful: nmap
Now we can move forward and backward in the hyperlinks in the help system, with easy to remember key bindings. Now let's also add some easy navigation keys for finding the next/previous place where a subject or an option link is situated in the currently open help file. This way we can easily scroll through the help file until we find what we are looking for. nmap nmap nmap nmap
o O s S
/''[a-z]\{2,\}'' ?''[a-z]\{2,\}'' /\|\S\+\| ?\|\S\+\|
Now you can press o to go to next place where an option link is, or s if you want to go to the next subject link. The same is available if you want to move backwards, you just have to press the capital letters instead—hence O for previous option link and S for previous subject link. To prevent the mappings from interfering with other key mappings, you can add them to a file called help.vim and place it in $VIMHOME/ftplugin/.
So now, we are only missing a final bit of our improved help system navigation. We need a way to open the help system a bit faster. Normally, when you press the F1 key, the help system opens on the default page. It would, however, be nice if the key instead did a lookup of the word currently under the cursor. So let's look at a key mapping for this: :map <ESC>:exec "help".expand("")
This one is a bit hard. As the :help command is normally used for looking up Vim commands, the commands on the line after the :help are not interpreted. Because of this we have to wrap the command in the :exec command.
[ 56 ]
Chapter 3
To get the word under the cursor, we use . The WORD part is in uppercase; this means that all characters except white spaces (space and tab) can be part of the word. This is needed because Vim commands can contain special characters other than alphanumeric (think of if you were to look up ). This key mapping can be used from outside the help system, and could therefore be added to your vimrc file, and not placed in help.vim.
Faster Navigation in Multiple Buffers
Often, you are not just working on one file, but have multiple files open. For every file you have open, you have a Vim buffer. A buffer can be shown or hidden, which means that to find the file you want to work on, you will need to find the buffer containing it. You could of course, bring up the list of buffers and find the right buffer in the list. To show the list of buffers you can use the command: :buffers
This list is not interactive, so in order to select the buffer you want to go to, you need to look up the number at the beginning of the line where the file is listed. This is the number of the buffer where the file is placed. With this number, you can now go directly to the buffer by executing the following command: :buffer N
where N is the number of the buffer. This way of navigating the buffers is not always the most efficient. You could also cycle through the buffers by using the following commands: :bnext :bprevious
Even though these commands can be accessed via their shorter names :bn and :bp, they are still commands you have to write in normal mode. This means that it takes at least five key presses to execute the command, which is not convenient. So in order to make this buffer cycling a lot faster, you could add the following mapping to your vimrc file: map <ESC>:bn map <ESC>:bp [ 57 ]
Better Navigation
What these map lines do is make it possible to use Ctrl+left arrow key to go to previous buffer and Ctrl+right arrow key to go to next buffer. So by holding down Ctrl while pressing the left/right keys repetitively, you can easily and quickly cycle through the files you have open. If you want to toggle back and forth between current and previous buffer, then you can use Ctrl-6 (Ctrl-o Ctrl-6 if in insert mode) or :e #.
Open Referenced Files Faster
In many programming languages, you can include other files in the current file, and thereby split the contents across multiple files. Often the inclusion of the file resembles something like: #include "somefile.h"
Here we have "somefile.h" as the name of the file we included. It would be nice to have an easy way to open the included file. Vim has a command that helps you in doing exactly that. Move the cursor to the place in the file where the filename of the file you want to open is, and execute the following command in normal mode: gf
You can remember this command by thinking of 'goto file'. Vim looks for the file in several different places: 1. Vim looks in the places it has defined in the Path option, and relative to the currently open file. 2. If not found, Vim uses the suffixadd function to see if it can find the file by adding one of the suffixes (e.g. adding .c to the filename). 3. If still not found, Vim uses the includeexpr expression to convert the filename to something that is hopefully understandable as a filename (for example, java.com.http is translated to java/com/http.java)
[ 58 ]
Chapter 3
If Vim finds the file, then it opens the file in the current buffer, and if not, it returns an error message. If the buffer you are currently in is not saved, or if anything else is going on such that Vim cannot abandon the currently open file, then Vim cannot open the file. This can be quite annoying, but it is a problem we can prevent from happening. By simply adding the following command to your vimrc file, you will always open the new file in another buffer and Vim does not have to abandon the currently open file: :map gf :edit
This command simply overwrites the gf command to instead open the file under the cursor with the :edit command—and if it does not exist, then open a new empty buffer. If you want Vim to support filenames with spaces in when using gf, then add the following to your vimrc: set isfname+=32 32 is the decimal number representing space in the ASCII table.
Search and You Will Find
We all know the feeling of having seen the things we have misplaced somewhere, but not remembering exactly where. What we normally do in a situation like this is search for the thing we are missing. In Vim, we can do the exact same thing. Let's split the search into three cases: 1. Searches in current file 2. Searches in multiple files 3. Searches in help files In the following sections, we will look at recipes that help you with the three types of search.
Search the Current File
Even though your file might not be that long, it can still be a pain to find something you are looking for. Vim has several ways to help you find what you are looking for. So let's look at some examples.
[ 59 ]
Better Navigation
Example 1—Find next occurrence of a word You know that around where you use the word "someWord", you have the text you are looking for. To find this, you simply need to do a search for it by executing the following command in normal mode: ?someWord
The command searches backwards in the file for the first occurrence of the word after the question mark. If you are at the end of the file, this is the perfect way to search for a word, but if you were at the beginning of the file it would make more sense to search forward in the file. This is done by exchanging the question mark for a slash: /someWord
The word might be in the file several times, and maybe the first place you found wasn't the place you were looking for. No worries, you simply need to press n to go to the next occurrence of the word in the direction of the search. If you would rather change direction, then simply press N instead, and it will instead find the preceding occurrence of the word. If you want to do the same search again, simply use ?? or // instead of writing the entire word again. If you add set incsearch, your search will be live and the cursor will start jumping through the file as you type. At any point while writing, the cursor will be placed at the next/previous occurrence of the word as it is written right now. You have to press Enter in order to actually execute the search in the end, else the cursor will go back to where it came from. To cancel a search and go back to where you came from, you can simply press the Esc key.
Example 2—Search for word under cursor If you are already near one occurrence of the word you are looking for, but it is just not the right one, or maybe you want to look through all places where a certain word is used, and the word is already written, why use extra key presses on writing the word again? Vim has just the right commands for you. Place the cursor on or just in front of the word you want to search for and press either one of the following two keys in normal mode: # *
[ 60 ]
Chapter 3
The first one searches for the previous occurrence of the word under the cursor, and the second one searches for the next occurrence of the word. Pressing the key multiple times jumps to the next/previous occurrence of the word, over and over again. This makes it really fast to jump through all occurrences of the word. Maybe your word isn't actually a complete word, but just a part of a word. Vim also has a command for this. Simply press the following key combination in normal mode: g# g*
Now Vim does not just jump to the next occurrence of the word, but also to any occurrences where the word is part of another word. For example, placing the cursor on the word "foo" and pressing g# will make Vim jump to the next "foo", in both "foobar" and "food."
Search in Multiple Files
Maybe what you are looking for is not in the current file. Maybe you are not even sure which file you should be looking in to find what you are looking for. On a Unix-flavored operating system like GNU/Linux, you typically have the command-line tool grep, that looks for certain words or patters in all the files specified. In Microsoft Windows there is a similar tool available as the commands FIND and FINDSTR. These are, however, not commonly used by Windows users. In order to provide all Vim users, no matter which platform, with a way to search through files, Vim has its very own grep command. The command to use is: :vimgrep /pattern/[j][g] file file2... fileN
This command takes two arguments. The first is the pattern you want to search for. You can use Vim's regular expressions in the pattern or you can just write a word. The pattern needs to be enclosed in / and after the last / you can add either of the two flags j and g. The flags helps you select how much to get in your result, and how it should be presented to you. Instead of / around your pattern, you can use any non-ID character. A non-ID character is any character not defined in the isindent option.
[ 61 ]
Better Navigation
If the g flag is added, then the result will include a line for each match of the pattern. This means that if your pattern is matched three times in the same line, then you will get the line three times in your result. If the j flag is added to the end of your pattern, then you will not be presented with the result but it will just be updated into your quickfix list for later retrieval (see :help quickfix for more information about quickfix lists). Without the j flag, you will be moved directly to the first match and the rest of the result will be added to your quickfix list. To show your quickfix list with the vimgrep result in, simply use the command: :clist or navigate to next/previous match with :cnext/:cprevious
The second argument to the Vim grep command is the list of files you want to search through. The file list can consist of a single filename, a list of filenames or a pattern using the star wildcard (for example, *.c *.h). You can also use the ** wildcard, like **/*.c if you want to search in all the C files in the current folder, and recursively through all subdirectories.
Search the Help System
Sometimes when you need help for something in Vim, you might not know exactly what you should look for. You could of course start going through the entire help system, but it consists of several different files and thousands of possible keywords. So, Vim has the right command to help you out here. As in the previous recipe, the keyword is grep, and for the Vim help system it is centered around the following command: :helpgrep pattern [@LANG] [ 62 ]
Chapter 3
The command takes one argument, the pattern you search for, plus one optional argument to limit the language. Let's look at an example to make it clearer. You need some information on auto completion but do not know where to look for it. You are able to read English hence only want help in this language. A search for this could look like this: :helpgrep completion@en
What the command does is search for the word completion through all the English (en) documentation. The command takes you to the first match it finds and the rest of the matches are added to the quickfix list for later retrieval. If you want to use the location list instead of the quickfix list for your result, then you can use the command :lhelpgrep instead.
The helpgrep command does not actually look through all the documentation when searching, but uses a tag list containing tags for all the available documentation to look up a pattern. This tag list is, however, not created automatically, so it is important to note that if you install a Vim plugin that has its own documentation, then you need to use the following command: :helptags /path/to/documentation
The path to the documentation only needs to be where you have installed the new documentation. But in order for the Vim to actually be able to find the documentation, it has to be in a docs/ directory in one of the places defined in the runtimepath in Vim (see :help 'runtimepath')
X Marks the Spot
Sometimes when editing a line in a file, you have to go to somewhere else in the file to look up something. Afterwards, it can be difficult to find the line you were editing, and you waste valuable time on finding it. Wouldn't it be nice if you could mark the spot before leaving it, such that it is easy to find later? Vim has some tools for you that can do just that. We can split it into two categories: •
Visible markers
•
Hidden markers
[ 63 ]
Better Navigation
In the following two sections, we will look at the possible ways of adding marks in Vim, and then it's up to you to figure out which one fits your needs the best.
Visible Markers—Using Signs
In Vim we have a nice feature for marking a line with a visible mark—signs. A sign is a mark that will show up in the leftmost column in the editor. If you want to change the color of the column in which the sign is shown, then you can use the following command: :highlight SignColumn guibg=darkgrey
Depending of whether you are using Vim in a console or as Gvim, the sign can be either a combination of characters (for example, >> ) or an icon. To use the signs you will need a bit of setting up. You only have to do this once if you have it in your vimrc file. The first thing you have to do is define the signs you want to have. The command you need to use is as follows: :sign define name arguments
The arguments can be one of the following: •
linehl : Color group you want to mark the line with.
•
text
•
texthl : Color group you want the sign text marked with.
•
icon
: The text used as a sign in console Vim (e.g. >> !! or ++). A maximum of two characters can be used per sign. : Full path to the icon you want for the sign in Gvim. The icon should be small enough to fit the size of only two characters. The format should be a bitmap format, but .xpm format is preferred.
An example could be: :sign define information text=!> linehl=Warning texthl=Error icon=/ path/to/information.xpm
Now we have defined a sign and added it to our vimrc file, and are ready to place the sign somewhere. The command is : :exe ":sign place 123 line=" . line(.) ."name=information file=" . expand("%:p")
Replace the number 123 with any number you will use as ID for this sign. [ 64 ]
Chapter 3
As you can see, this is a bit harder, but it can easily be mapped to a key. What it does is add the sign named information under the ID 123 to the current line (line(.)) in the currently open file (expand("%:p")). Mapping this to a line is: :map :exe ":sign place 123 line=" . line(".") ."name=information file=" . expand("%:p")
This maps the information sign to the F7 key such that it will be placed in the current line whenever you press the F7 key.
Sometimes we also want to remove the sign again. In Vim this is called to 'unplace' a sign: :sign unplace ID
The ID is the ID you gave your sign when you placed it (123 in the above example). This removes the sign from all the places where you have added the sign with that ID. You might want to remove it only from the current file, and can therefore add another argument for the file like this: :sign unplace ID file=name
Or from the buffer: :sign unplace ID buffer=bufferno.
where bufferno is the number of the current buffer (see :buffers). If you want to remove the sign in the current line, then you can simply use: :sign unplace
Let's map this to Ctrl-F7, just to make it symmetric with the sign placement mapping we have defined earlier: :map :sign unplace [ 65 ]
Better Navigation
If you have added several signs with the same ID to a file, then the above mapping will only remove the uppermost sign with the specific ID and not the one in current line.
As this is a chapter about navigation, we also need to have a bit about navigating to a sign here. This is called 'sign-jumping' in Vim and uses the following command: :sign jump ID file=file
Here ID is the ID of the sign you want to jump to and file is the file you want to find the sign in. Instead of file=file, you can instead use buffer=bufferno. Again, if the sign has been added with the same ID several times in the file/buffer, then it will jump to the first sign in the file. Poul Rouget has created a Vim script that makes the usage of signs a lot easier. You can find it here: http://www.vim.org/scripts/script.php?script_id=1580
Hidden Markers—Using Marks
Marks is the fast and easy way to add a mark to the current line such that you can later jump to it easily. Basically, it consists of a normal mode command that sets the mark, and a normal mode command to jump to the mark. You won't be able to see if a line is marked or not unless you open the list of marks. So let's look at how to mark the current line. We simply press the key m in normal mode, followed by one of the characters 0-9, a-z, or A-Z. If you for instance press ma, then it means that the current line is marked with the mark named a. If you later want to jump to this line, then you simply press 'a (single quote + mark name) and you will be taken to the beginning of the line you marked (if indented, then to just before the first non-whitespace character). In some cases it might not be efficient to be placed at the beginning of the line, but it would be much better to be placed where you were when you added the mark. To jump to this place instead, you simply replace the single quote with a ` (backtick) like `a.
[ 66 ]
Chapter 3
The different mark names have different meanings and work areas: •
0-9 : Marks set from .viminfo and normally only used by Vim itself (e.g. mark 0 is the place where cursor was when the file was last exited). A user can, however use this to make an "open recently used" functionality.
•
a-z : Marks only available in the current file. These marks are deleted when the file is closed. You can only jump to a lowercase mark if you are inside the buffer containing the file.
•
A-Z : Marks available across files. These marks can be jumped to, even if you are not in the file where the mark is situated. If a viminfo file is available then these marks are saved until next time you edit a file.
You can always get a complete list of your marks by using the following command: :marks
This shows which files the different marks are set in and on what lines. To delete one or more marks, you can use the command: :delmarks markid markid...markid
Examples of how it can be used are: :delmarks a b c :delmarks a-c :delmarks a f-i 1-4
If you want to delete all marks in the current buffer, then simply use the command: :delmarks!
Other types of marks are set by Vim all the time when using it. These can be marks for where the cursor was last time the insert mode was exited, beginning/end of text selected in visual mode, the last place you changed something, etc. Look in :help mark-motions for more information on how to use marks and which other types of marks you have available.
[ 67 ]
Better Navigation
Summary
In this chapter, we have looked at alternative ways for boosting the speed at which we navigate through files and buffers in Vim. First, we looked at how to navigate through a single file faster by using the contextual structure of the file for navigation. We also looked at a nice recipe for how we can make it easier to navigate files with long, wrapped lines. Next, we looked at how to to navigate the Vim help system faster and learned how simple key bindings can make the help navigation more intuitive and recognizable. Now, we knew how to navigate inside a file, but we also needed to know how to navigate between files and buffers. The next section took us through how to navigate the buffers faster and how to open a file that is referenced by another file with only two key presses. We can navigate in many ways, and in the preceeding sections we looked at how to use the search mechanisms in Vim to navigate not only the open files, but also files on the disk. We also learned how to use searches in the help system to find help on topics we could not find normally. Finally we have looked at how to use signs and marks to jump around in files, and how Vim helps us simply by adding some marks automatically when we use it.
[ 68 ]
Production Boosters In this chapter, we will look at how even small changes can make work go faster and more smoothly in Vim. Some recipes introduce you to features in Vim, while others will show you how scripts can help. It doesn't matter whether you use Vim for making small changes to configuration files, or if you use it as you primary editor in a large development project, you will find recipes in this chapter that can help you improve your performance when using Vim. This chapter contains recipes that cover: 1. Templates using simple template files 2. Templates using abbreviations 3. Auto-completion using known words and tag lists 4. Auto-completion using omni-completion 5. Vim macros and macro recording 6. Using sessions 7. Project management using sessions 8. Registers and undo branches 9. Folding for better overview and outlining 10. vimdiff for change-tracking 11. Opening files everywhere using Netrw After reading this chapter, you should be able to boost your productivity in Vim by several percent.
Production Boosters
Using Templates
No matter what type of files you are working with, there are always some basic things to set up when the starting off on a new file. Creating this setup is a tedious task, and even worse is the fact that you have to do it again when you start on a new file. So why spend a lot of time on these things when you could just as well create templates for these types of structural patterns? In the next couple of sections, we will look at recipes for two types of templates: 1. File type-specific templates for new files 2. Content-pattern templates So let's get started on creating some templates.
Using Template Files
Every time you start working on a new file, it is most likely that the first thing you'll do is add some sort of header (or other information) to the file. What you have to add is of course, dependent on which file type you are working on. Some examples could be : • • •
Adding basic structure (, , and ) to new HTML files Adding a header to all C files and also a main function to main.c files Adding the main class structure to a Java file
You can probably find many other things you would like to add to the file types you work with. So, how do we create a template file? Let’s use an HTML file template as an example. The structure in such a file is quite static, and hence great to have a template for. Our simple template could look like: <meta name="generator" content="Vim" /> <meta name="author" content="Kim Schulz"/>
Content goes here...
[ 70 ]
Chapter 4
We create a directory in our VIMHOME called templates/ and place a file with the above HTML code in the directory—save the file as html.tpl. Now the first template is in place, but we need to get it loaded into all new HTML files that we create. To do so, we add the following auto-command to our vimrc file: :autocmd BufNewFile *.html 0r $VIMHOME/templates/html.tpl
What this command does is ensure that when you create a new file with the file extension *.html, the content of your template file is read into your new file. This way your file gets prepared with the template's content before you can start editing it. All this is very nice, but after adding a bunch of templates, you might get tired of adding lines to your vimrc file. So let's make our first line a bit more intelligent: :autocmd BufNewFile * silent! 0r $VIMHOME/templates/%:e.tpl
What this single line does is that, whenever you open a file, it looks for a template that matches the extension of the file—for example, when creating the file index.html, it looks in $VIMHOME/templates/ for a file named html.tpl. If there is no template for the file type, then it simply creates an empty file as usual. Let's take these templates even further by adding support for placeholders (for places where you want to add text to the file, fast). A placeholder could look very different depending on what you like, but I propose something like . So if we take a line from the HTML template mentioned above, and add a placeholder to it, it could look like: <meta name="generator" content="" /> <meta name="author" content=""/>
[ 71 ]
Production Boosters
Now we have the placeholders in place and only need a way to jump between them. So let's add a command to our vimrc that will make it easy to make this jump. We want to use Ctrl-j as the jump key binding, because it can easily be used in the insert mode and the 'j' (for jump) makes it easy to remember. The command could look like this: nnoremap /c/+>/e inoremap <ESC>/c/+>/e
Now you can easily jump to the next placeholder in the file, change the text, and jump on to the next placeholder—simply by pressing Ctrl-j text, Ctrl-j text, etc. By having the keyword in the placeholder, you can easily see what you are supposed to add there. You can mark your placeholders by adding a match command to your vimrc: match Todo // (replace Todo with whatever color group you like).
Abbreviations as Templates
In the previous section, we've learned how to make templates for entire file types, so now, let's look at how to make templates for patterns inside the file-content itself. In Chapter 2, we briefly looked at how to use abbreviations for limiting the amount of key presses, whenever possible. Now, let's take the idea of using abbreviations and copy it to our template system. Let's look at the command and what it's all about, just to refresh our memory: :iabbrev match replace-string
We only want the command for insert mode, because it is there we want to use the pattern templates. An example could be the following pattern template for a C file: :iabbrev for( for (x=0;x at the beginning of the line, is the tag you are currently at. When using Ctrl-] and Ctrl-t, you move up and down in the stack, but you can also work the stack with commands: :tag move to next tag in stack :pop move to previous tag in stack
[ 76 ]
Chapter 4
When you have used a lot of tags and jumped around between them, then it might be nice to get a list of the ones you have available. To get such a list, simply use one of the following commands: :tselect :ptselect
The first one gives you a list of matching tags and you can then select the one you want by pressing the number you find at the beginning of the line. The second command does the same, except that it shows the list in a preview window. If you selected the wrong tag in the list, or just want to see one of the others, then you can move between them with the following commands: :tnext : Move to next tag in list :tprev : Move to previous tag in list
You might not see the strength of the tag lists in this example, but imagine that you project was not just three files, but 1000 files across hundreds of directories. Then you suddenly just can't go around remembering where each function is, and you would need a thorough indexing mechanism like tags files.
Easier Taglist Navigation
On most non-english keyboard layouts, the ] key is not directly available and you would need to press, for example, Ctrl-AltGr-9 to execute the Ctrl-] tag jump. In those cases, it would be nice to map the commands to more accessible keys. I use the following mappings: :nmap :nmap :nmap :nmap :nmap
<S-F7> :ptselect :tnext :tprev
Now, you can jump forth and back between tags with F7 and Ctrl-F7, get a list of the tags with Alt-F7, and go through the used tags with F8 and Ctrl-F8.
Other Usages of Taglists
Tag lists are not only used by programmers for looking up functions and variable definitions. They are also used for a lot of other interesting things. Just to give you an idea, here's a short list of examples where they is used in Vim scripts:
[ 77 ]
Production Boosters
•
lookupfile.vim Hari Krishna Dara has created a script that can use tag lists to find files in a tagged project simply by writing the filename. Find the latest version of the script here: http://www.vim.org/scripts/script.php?script_id=1581.
•
taglist.vim Yegappan Lakshmanan has created the taglist plugin for Vim, which is very popular among programmers. It is a complete source-code browser that gives a great overview of the functions, keywords, variables, definitions, etc. in a split-window. In Gvim it can even give you a complete menu with the tags of the project. You can find a lot more information about this plugin here: http://www.vim.org/scripts/script.php?script_id=1581.
•
ctags.vim Gary Johnson and Alexey Marinichev have created a simple, yet powerful plugin called ctags.vim. It simply shows the name of the function the cursor is currently placed in, in the status bar or window title. The script automatically generates tags files for the currently opened file using the program Exuberant Ctags. Find more info about the script here: http://www.vim.org/scripts/script.php?script_id=610.
•
autoproto.vim Jochen Baier has created a very useful script for C programmers. This script shows the prototype of the currently typed in function in a preview window whenever the programmer presses the first ( after the function name. Find more info about the script here: http://www.vim.org/scripts/script. php?script_id=1553. You can read a lot more about tags and how to use them in the Vim help system under :help tags.
Using Auto-Completion
As a Vim user that obeys the philosophy of Vim, you will do anything to minimize the number of key presses—because extra keys pressed equals extra time wasted. So why type each word to the end, when Vim is able to guess what you are typing and automatically complete the word for you? In Vim, there are multiple ways to auto-complete the words you are typing. Some methods simply complete words you have written once in one of the opened buffers, while others involve analyzing the code you are working on—not just the current file, but the entire source tree. [ 78 ]
Chapter 4
In the following sections, we will look at three different ways to use auto-completion in Vim: •
Auto-completion with known words
•
Auto-completion using a dictionary file
•
Context-aware auto-completion with Omni-completion
There will also be some small tricks on how to make it more comfortable to use auto-completion by using well recognizable key bindings.
Auto-Completion ��������������������� with Known Words
In this recipe, we will look at maybe the simplest type of auto-completion and at the same time the most overlooked—auto-completion with known words. Almost no matter what you are writing, you will eventually write the same words over and over again. In Vim, you can simply type in the first couple of letters of the word and then press Ctrl-n. Example: You want to write the sentence: "I have beautiful flowers in my flower garden." Since you have no text besides this in the file, you will have to write the entire first part of the text until it looks like: "I have beautiful flowers in my f" Now, you would normally continue the word "flower" but since it is already there, you can simply press Ctrl-n and the word will expand to "flower". As your text evolves, you will see that you can start using auto-completion on more and more words. What Ctrl-n actually does is look for a matching word by going forward through the file. If you know that you have just used the word, then it will be faster to use Ctrl-p instead because Vim will then search backwards in the file for a matching word. In general, you won't feel the difference unless you are working with really large files, or there are many possible matches.
[ 79 ]
Production Boosters
Auto-Completion ��������������������������������� using Dictionary Lookup
A neat trick is to find a large dictionary file with all kinds of words in your favorite language, then load this file into Vim as an dictionary (such files can easily be found on the Internet). To load the file into Vim as a dictionary, simply add it to the dictionary setting with: :set dictionary+=/path/to/dictionary/file/with/words
Now Vim suddenly knows a lot of words beforehand, and you can simple auto-complete using these words. Something is different, however. Now the words we use to look in are not words from one of the open buffers, but keywords from one of the dictionary files available in the dictionary setting. This is why you will need to use another key binding in order to do the completion this time. ctrl-x ctrl-k
By pressing Ctrl-x you get into a completion mode, and by pressing Ctrl-k you do a lookup for a keyword (remember k for keyword) in the dictionaries.
Other completion types are available. Some of them are in the following list: Ctrl-x plus: •
Ctrl-l Complete whole lines of text
•
Ctrl-n Complete words from current buffer
•
Ctrl-k Complete words from dictionaries [ 80 ]
Chapter 4
•
Ctrl-t Complete words from thesaurus (see :help 'thesaurus')
•
Ctrl-i Words from current and included files
•
s
Spelling suggestions (Vim 7.0 and newer only)
Others will be further described in the next sections.
Omni-Completion
We all have our perfect solution for what should be auto-completable and what shouldn't. In Vim there had been no way to give the user complete control over what to do about completion until version 7.0 came out. Vim 7.0 introduced a new completion technique called 'omni-completion'. It gave the user the possibility to define exactly how the functionality of the completion should work—in fact, then, the user would have to write the completion function himself or herself (unless someone else has already done it). As with the completions mentioned in the previous section, the completion is invoked by typing in some letters and then going into completion mode by pressing Ctrl-x followed by Ctrl-o to make an omni-completion. To add your own completion function, you simply do the following: :set omnifunc=MyCompleteFunction
Now you would just have to create a function called MyCompleteFunction that gives you the completions. This setting is only available to the currently active buffer, and you will have to set it for all buffers where you want to use it. Setting omnifunc is normally done in a file type plugin such that it is bound to a single file type.
So let's look at an example on how such a function could look like. If for instance, you have a file with all your contacts with one name + email address on each line like: Kim Schulz|
[email protected] John Doe|
[email protected] Jane Dame|
[email protected] Johannes Burg|
[email protected] Kimberly B. Schwartz|
[email protected] ...
[ 81 ]
Production Boosters
Now, you would like to insert an email address by writing a name and doing auto-completion on it. A function for this could look like: function! CompleteEmails(findstart, base) if a:findstart " locate the start of the word let line = getline('.') let start = col('.') - 1 while start > 0 && line[start - 1] =~ '\a' let start -= 1 endwhile return start else " find contact names matching with "a:base" let res = [] " we read contactlist file and sort the result for m in sort(readfile('/home/kim/.vim/contacts.txt')) if m =~ '^' . a:base let contactinfo = split(m, '|') " show names in list, but insert email address call add(res, {'word': contactinfo[1], \ 'abbr': contactinfo[0].' ', \ 'icase': 1} ) endif endfor return res endif endfunction
The function takes two arguments, which are needed for an omni-completion function. The first time Vim calls the function, it sets the first argument, findstart, to 1 (and base is empty) which means that this is the first invocation, and that it should find the beginning of the word you have written so far. Vim then invokes the function again, and this time with findstart set to 0 and base set to the word you have started auto-complete on. This time the function opens up the contact list file and reads it line by line into a list. It sorts the list such that the list in the completion popup is ordered, and then it iterates over the list. The lines are split at the |, and then the ones that begin with the same letters as the word you completed on are added to a result that the function will return. The function can modify how the popup content looks and what it should match on. This is done by not just adding the email addresses, but instead building a dictionary (see :help Dictionary) where some specific keys are set. In this case, we use three keywords: [ 82 ]
Chapter 4
•
word The actual word that should be inserted.
•
abbr
This word is used instead of "word" in the popup list.
•
icase
If this is a non-zero value, then the matching is case-insensitive.
Other keywords and their functionality can be found in the help system under :help 'omnifunc'
So now Vim has a list of words for its popup, or in this case small lines like: "Kim Schulz "
Whenever you write some letters like "ki" and then pres Ctrl-x Ctrl-o, then Vim will show the popup with all the names that starts with "ki".
To move between the items in the list, you can keep pressing Ctrl-o to cycle through the list. Alternatively, you can press Ctrl-n to move forward in the list and Ctrl-p to go backwards in the list.
A �ll-in-One Completion
You might wonder how you will ever be able to remember all these keyboard shortcuts and why you could not just use the same for all completion types depending on which one you have available. With Vim, you can of course do this if you want. So let's look at how we can do this in a way that is easy to remember for you. Nearly any other editor that supports completion has this functionality mapped to the tab key. In the help system, you will find a function called CleverTab() if you look in :help ins-completion
[ 83 ]
Production Boosters
This function lets you use Tab to complete words instead of Ctrl-n. It could distinguish between whether it should insert a tab character or do completion. If you pressed tab at the beginning of the line (indention) or after another whitespace character then it inserts a tab character—in the rest of the cases it would try to do known-word completion. We take this CleverTab function and extend it even further such that it selects the completion method to use from this prioritized list: 1. Omni-completion 2. Dictionary-completion 3. Known-word-completion A function that can do this, could look like the following: function! SuperCleverTab() “check if at beginning of line or after a space if strpart( getline('.'), 0, col('.')-1 ) =~ '^\s*$' return "\" else “ do we have omni completion available if &omnifunc != '' “use omni-completion 1. priority return "\\" elseif &dictionary != '' “ no omni completion, try dictionary completio return “\” else “use omni completion or dictionary completion “use known-word completion return "\" endif endif endfunction “ bind function to the tab key inoremap =SuperCleverTab()
Add the function and the binding to your vimrc file, and then you are ready to do completion with your tab key. You simply have to press the Tab key to do you completion, and the function checks to see if it should insert a tab character. If not, then it checks to see if you have a omni-completion function (in omnifunc) available. If this is also not the case, then it looks if there is a dictionary available. If you have no dictionaries available, then it falls back on using simple known-word completion. [ 84 ]
Chapter 4
Using Macro Recording
Probably the most overseen production-booster when working with monotonic structured text is the ability to record input macros to do them over and over again. The interface for doing this is extremely simple, but nearly everything can be recorded, so it reveals a very powerful tool. Let’s start by looking at the commands to use: qa : Record from now on into register a. Any register can be used but q is often used
for simplicity.
q
: If pressed while recording, then the recording is ended.
@a : Execute the recording in register a (replace with any register). @@ : Repeat last executed command.
You can add any number before the @ to repeat the execution of the recording that number of times. For example 15@a will execute the recording in register a 15 times. So let's look at a normal recording session in Vim: qq command1 command2 .... commandN q 10@q
You might wonder what this can be used for; because when is it exactly that you need to execute a list of commands over and over again? This is best shown with an example. Imagine you have a large list with information. It could e.g. be a log file from a Unix system. That could look something like: Oct Oct Oct Oct Oct Oct Oct Oct Oct
8 8 8 8 8 8 8 8 8
21:23:34 21:23:34 21:23:34 21:23:34 21:23:34 21:23:34 21:23:34 21:23:34 21:23:34
laptopia laptopia laptopia laptopia laptopia laptopia laptopia laptopia laptopia
kernel: kernel: kernel: kernel: kernel: kernel: kernel: kernel: kernel:
ACPI: bus type pci registered PCI: PCI BIOS revision 2.10 entry at 0xe9694 Setting up standard PCI resources ACPI: Subsystem revision 20060127 ACPI: Interpreter enabled ACPI: Using IOAPIC for interrupt routing ACPI: PCI Root Bridge [PCI0] (0000:00) PCI quirk: region 1000-107f claimed by ICH6 PCI quirk: region 1300-133f claimed by ICH6
[ 85 ]
Production Boosters Oct Oct
8 21:23:34 laptopia kernel: PCI: Ignoring BAR0-3 of IDE controller 001:1 8 21:23:34 laptopia kernel: PCI: Transparent bridge – 0000:00:1e.0
...
Now, you want to convert this file into an HTML file where the info is presented in a table with each column of text represented by a table column. You basically want each line to look like:
Oct 8 21:23:34 | laptopia | kernel: | ACPI ... |
You could start editing the file line by line, until you came to the end of the file. But how about just editing one line while recording it, and then playing back the commands on the other lines? The command execution could look like the following starting with the cursor placed at the beginning of the first line: qa
: Start recording into register a.
i|
[ESC] : Go into insert mode, insert first HTML, go back to normal mode. / [CR] : Search forward for whitespace. 3n : Advance 3 whitespace searches forward. xi | [ESC] : Delete whitespace, go into insertmode, add HTML, to normal mode. n : Goto next whitespace. xi | [ESC] : Delete whitespace, go into insertmode, add HTML, to normal mode. n : Goto next whitespace. xi | [ESC] : Delete whitespace, go into insertmode, add HTML, to normal mode. A |
[ESC]
: Append the final HTML to the end of the line and go in normal mode.
j^
: Advance a line and move to the beginning of it.
q
: End macro recording.
[ESC] press escape key, [CR] = press return key So now we have one line ready, the cursor placed just correct on the next line, and a macro recording to play back on the rest of the lines. You can playback the recording with @a or you can simply play back the command on each line in the file with 9999@a. All you need now is to add a header and a footer to the file, but that is not really interesting here. This is just one place where macro recordings can be used, and if you think back, you will probably remember situations where you could have optimized your work by using a macro recording. [ 86 ]
Chapter 4
Using Sessions
Have you ever wondered how much information Vim actually holds for you about a wide range of settings and things like: •
Open files, buffers, windows, and tabs
•
Command history
•
Points of change in the text
•
Selections and undo branches
•
Size of windows, splits, and GUI window
•
Place of cursor
...and many other things. The stored information can be split into three different categories: The first type of setting is called a View and applies to a single window in Vim. A view can be saved and restored such that a window will have the same look and setup every time you use the view. The second type of settings are called Sessions and are collections of views and info about how they are inter-operating. Like views, sessions can also be saved for later retrieval. The final type of setting includes all the rest; all the global settings that does not directly apply to any window in Vim. These settings are stored with the session, such that they can also be saved/restored. In the following sections, we will look at how sessions can be used for different tasks during your daily work in Vim.
Simple �������������������� Session Usage
When using sessions, the most basic thing to do is to save the currently running session (default session when no special session is loaded) to a session file, such that you can load it again later when you need to. The main command to use is: :mksession FILE
[ 87 ]
Production Boosters
or if you only want the save the current view, then: :mkview FILE
FILE is the name of the file you want to save your session or view to. If no filename is given, Vim uses a file called Session.vim, which it puts in the current work directory. If you have previously saved a session with the same filename, you can add a ! after mksession to make it overwrite the file.
When working with views, you can have many different views at the same time. If every view were saved in the current working directory, then it would be filled up with view files. To prevent this you can tell Vim where it should place the view files with the following command: :set viewdir=$HOME/.vim/views
In the above case, you will set it to store the view files in a directory called views, which is placed in your $HOME/.vim/ directory. So as an example, we could say that you have three windows open and just before closing Vim you do: :mksession
Then the next time you want to open Vim with the same session, you simply start Vim with the command-line argument -S, e.g.: vim -S Session.vim
Now Vim will be started with the same settings as when you saved the session. Alternatively, you can open Vim as you normally would, and then use the following command to load the session file: :source Session.vim
In the case of views, instead you can use: :loadview View.vim
Loading a session can change the entire layout of the editor; loading a view will only change the layout of the active window.
[ 88 ]
Chapter 4
If you want Vim to remember settings like cursor placement and folds, when moving between multiple folders, then you can add the following to your vimrc file: set viewdir=$VIMHOME/views/ autocmd BufWinLeave * mkview autocmd BufWinEnter * silent loadview
A view of the buffer is saved whenever you show another buffer in the same window, and restored when you show the buffer in the window again.
The trick is to add commands for saving a session when quitting Vim, and restoring the session when opening Vim. This way you can open and close Vim without losing the settings, list of open files, etc. You can do this by simply adding the following commands to your vimrc file: autocmd VimEnter * call LoadSession() autocmd VimLeave * call SaveSession() function! SaveSession() execute 'mksession! $HOME/.vim/sessions/session.vim' endfunction function! LoadSession() if argc() == 0 execute 'source $HOME/.vim/sessions/session.vim' endif endfunction
If you now close Vim, then it saves a session file in $HOME/.vim/sessions/session.vim. Depending on how you open Vim, it either opens the file specified on the command-line or reopens the latest session, for example: vim file.txt : This opens Vim without loading the last session vim
: This opens Vim with the last session loaded. Previously opened files are reopened.
If you want to store additional settings besides what Vim stores in the session file, you can add an extra session file. This is done by creating a file named like your session file, except that the .vim extension is replaced by x.vim. For example, Session.vim has the extra session file Sessionx.vim. The extra session file should be placed in the same folder as the session file itself. You can then add all the Vim commands you want to in this file, and these will be executed once the session file has been read. [ 89 ]
Production Boosters
Satisfy your own Session Needs
It is not always that you want everything saved in your session. Sometimes, it might just be the files you had open that you want to save info about. Other times you want to store every single piece of information you can about a session. Fortunately, Vim gives you a way to set up what you want it to save in a session file. The setting you should work with is called sessionoptions and can be set with: :set sessionoptions=OPTIONS OPTIONS is a comma-separated list with one or more of the following options:
•
blank
Save empty windows.
•
buffers
Save info about all buffers including hidden and unloaded buffers.
•
curdir
Save information about current work directory.
•
folds
Save information about folds in the buffer contents.
•
globals
Save information about global variables. Only variables starting with an uppercase letter and of the type String or Number will be saved.
•
help
Save the help window.
•
localoptions Save info about local options and mappings you have created for a single window.
•
options
Save all options, both local and global.
•
resize
Save info about the size of the UI window (lines and columns).
•
sesdir
If set, the current directory is the place where the session file is saved (cannot be used when curdir is also set).
•
slash
Change backslashes in all paths to slashes (make Windows paths Unix compatible).
•
tabpages
Save information about all tab pages and not only the active one, which is default without this option.
•
unix
Use Unix line endings, even on Windows systems.
•
winpos
Save information on where the UI Window was placed on the screen.
•
winsize
Save the size of all open windows.
[ 90 ]
Chapter 4
The options marked with bold are the ones Vim has turned on as per default. Instead of setting the entire list of options, whenever you want to add or remove a single one, you can instead use the '+=' and '-=' operators. If you, for example, have the default options but would like to have winpos added to the options, and also have folds removed, then simply do: :set sessionoptions+=winpos :set sessionoptions-=folds
You can see which options you have in your sessions options with: :echo &sessionoptions
You can in fact see any of the settings in Vim by simply using the :echo and adding an & in front of the setting's name, for example, :echo &somesetting
Sessions ����������������������������� as a Project Manager
You might sometimes want to use session files as a sort of primitive project file with information about some project you are working on. So when working on a project and having a lot of files and windows open, you simply use: mksession!
to save the current session to the Session.vim file in the folder that you are working from. It would then be ideal if Vim automatically loaded the session for the project if there was a project file (Session.vim). So why not make this possible? Simply add the following to your vimrc file: silent source! Session.vim
Now if there is a Session.vim file in the folder where you start Vim, then it will load it. So as long as you keep the session file in the project directory, you can easily reload the project in Vim over and over again—just remember to save the the session again if you open new files or change the windows/buffers.
[ 91 ]
Production Boosters
This is just a simple way of using sessions as a project manager, but it can be made a lot more advanced. Wenzhi Liang has created a practical script that adds a Projects menu item to Gvim. In this menu, it is possible to save the current session as a named project. Later, you can restore the project (which is now available directly in the menu) or you can switch between projects with a single click of a menu item. If you don't need a project anymore, then you simply choose to delete it via the menu.
You can find the latest version of the script and read more about it on this homepage: http://www.vim.org/scripts/script.php?script_id=279
Note that the script demands that you have both Perl and Bash installed on your computer.
Registers and Undo Branching
You might know the feeling when you delete or cut something from your text, and realize later that you needed that text elsewhere. If you have already copied/ cut another piece of text, then normally the old text is gone because the clipboard normally only has room for one piece of text—not in Vim. In Vim, you have two tools that can help you when you modify your text and need to keep track of deleted/copied text or changes to the text in general: •
Registers Registers is sort of an advanced clipboard with multiple buffers for storing your clippings, deleted, and copied text.
•
Undo Branching Undo Branching is a simple form of version control built into Vim. It gives you the possibility to roll back changes to a file until a certain time or number of changes. If you later regret undoing some changes, you can go back and find an undo branch containing those exact changes. [ 92 ]
Chapter 4
The next two sections will tell you more about how to use both registers and undo branching in your daily work. After reading them, you will easily see why these tools are very strong tools and how they can help you in your daily work in Vim.
Using ��������������� Registers
In many programs and operating systems, you only have access to a single clipboard for text you cut or copy. This is not the case with Vim, because here you have access to not one, not two, but nine different clipboards—or register types as they are called. Some of the register types overlay each other's working area while others have a very unique purpose. You can use the registers in connection with a range of commands and movements like yank, delete, and paste. The registers are all named with “ in front of the name like “x. So let's look at how to use a register. Let's just say that we use the register called “x in the examples. What x actually should be will be explained later. To store a piece of text that you want to copy into a register, you can use the normal y for yank, except that you start out by telling it where to yank it to: “x y (or "x yy if you want to copy the entire line)
the same is the case when cutting text with the x command: “x x
or when deleting text with the d command: “x d
So now you have the text stored in the register "x and you want to paste it again. You can simply use the p (before cursor) and P (after cursor) commands to paste the text and just start out by telling Vim which register to paste: “���� x p or “��� ���� x P
If you have forgotten which register you used, then you can simply type in the command: :registers
So now you know how to use the registers with the basic commands in Vim, and it is time to look a bit further at the different register types. The following sections describe each of the nine types of registers.
[ 93 ]
Production Boosters
The Unnamed Register
The unnamed register is called so because it is accessed via ““, hence resembling the empty string or no name. Vim automatically fills this register whenever some text is yanked with y or deleted with one of the following commands: d (delete), c (delete and go into insert mode), s (substitute), or x (cut). What this register does is point to the last used register, which also means that it will still work even if you use a specific register when deleting/yanking a text—for example “xdd will fill both register x and the unnamed register. If you paste some text with p or P without specifying any register, then it actually gets the text it pastes from the unnamed register.
The Small Delete Register
Whenever you delete less than one line of text, Vim will move it into a very specific register—the Small Delete Register (“-), only exception where this is not the case is when you specify another register to use.
The Numbered Registers
The numbered registers are named “0, “1, “2 and so on up to “9. They can be split into two types. The first type is register “0, which always contains the last deleted (d or x)/changed (c) text. When you delete or change something new, then register “0 is overwritten with the new text. Like register “0, register “1 also contains the last changed/deleted text. There is, however, the difference that register “1 will not be updated if another register is specified, or if the text is less than one line long (the small delete register is then used). For compatibility with vi, there is, however, an exception where register “1 is used, no matter what the length. This is if one of the following movement commands is used in your change/delete: %, (, ), {, }, `, /, ?, n, N Unlike register “0, the contents of register “1 is not deleted whenever new text is added to it. Instead, it is moved to register “2. If register “2 was already full, then this text is moved to register “3 first, and so on until register “9. The contents of register “9 will be overwritten whenever any new content is added to it. This way the registers “1 to “9 can function as a delete/change history such that you can get access to earlier deleted text, even if you have deleted new text more recently.
[ 94 ]
Chapter 4
The Named Registers
There are two types of named registers—“a to “z and “A to “Z. If you use the lowercase registers like “a, then they work like a normal register that you can copy deleted or changed data into. When new text is added to a register, then the old contents is discarded. If you instead use the uppercase registers (for example, “A) then the previous contents of the register is not deleted, but the new text is instead appended to the register. If you add the value '>' to your coptions, then the appended text in uppercase registers is split by a newline: :set coptions+='>'
Since you have the complete control over the named registers, they are most likely the type of registers you will get familiar with first.
The Read-Only Registers
There are four different read-only registers. What makes these registers so special is that only Vim has access to them. You only have access to pasting them with the normal P, p, or :put commands. The contents of the read-only registers are quite different: “% : This register always contains the name of the file in the currently active buffer. “# : This register always contains the name of the previous file in the currently
active buffer—also called the alternate file.
“.
: This register always contains the last inserted text. You will therefore be able to repeat the last inserted text by executing the normal mode command “.P.
“:
: This register contains the command you last executed on the command-line. If you repeat a command from the history, then this register will not be overwritten with the command. You will have to write at least one character of the command in order to get it stored in the register.
The Selection and Drop Registers
This register type consists of three registers: “*, “+, and “~. The registers are used to store and retrieve the text you have selected in Gvim. The “* register actually accesses the clipboard of your windowing system. If you use Microsoft Windows, then you wont feel any difference between using “* and “+. On Linux, however, there is a difference between the two registers because the Clipboard in X11 (the [ 95 ]
Production Boosters
windowing system) has not only one selection register, but three. The contents of the “+ register is any text you have selected. It is typically inserted by pressing the middle button on the mouse. The contents of the “* register is, however, only altered if you actually tell Vim to yank the text. These registers can be accessed from any GUI application, and are part of the normal copy-paste procedure you know from your daily work. The last register in this group is the drop register “~. This register contains the last selection that has been dropped into Vim. So, if you select some text in another application and drag it to your Gvim window to drop it there, then “~ will contain this text.
The Black Hole Register
As the name of this register indicates, this register works like a black hole—everything that goes into it never comes out again. This register is used if you want to completely delete some text and don't even want a record of it in any register. The black hole register is “- and can be used like for example “-x or “-dd. If you try to read out the text you have just written to this register, you will see that no matter what you do, it doesn't return anything.
Search Pattern Register
Whenever you do a search with the / command, the pattern you are searching for is automatically added to the search pattern register. The register name “/ is easy to remember because it resembles the search command / and just has the quotes added to show that it is in fact a register. Vim uses this register when you have hlsearch (highlight search pattern) turned on. You can use this to your advantage because, as this is a register, you can just change its contents to get hlsearch to highlight something different. To change the contents of the register without doing a new search, you can simply use: :let “/ = PATTERN
where PATTERN is what you want hlsearch to highlight.
The Expression Register
The expression register is the last register type in Vim. Calling it a register is, however, not exactly true because it does not store text as the normal registers do. You can not even write to it. Instead, it opens up the possibility to get access to the command line, execute an expression, and the get the result returned as if it were already stored in the register. [ 96 ]
Chapter 4
You get access to the expression register by simply typing its name “=. After pressing the equal sign, the cursor will be moved to the command line. You can see that you are working inside the expression register if there is an equal sign as the first character on the command line. You can now write the expression you want the result of, and then end by pressing the return key, or alternatively press Esc to return without executing the expression. If you press the return key without writing any expression, then Vim will find the latest expression executed and then use this instead. The expression needs to be valid and return a string. If the result of the expression is a number, then Vim automatically converts it into a string, but if you are unsure what type the result has, then simply use the string() function to convert the result before returning it. Look at :help expression to see how to put together a valid Vim expression.
Using �������������������� Undo Branching
We all know the feature where we can undo changes that we have done to a text. Vim has taken it a bit further and added the concept of branching to this. In this recipe, we will look at just what undo branching is, and how it can help you in your work. Let's start by defining what an undo branch is, in Vim. Let's say that you have a file where you have applied a range of changes. At some point, you realize that the last four changes were wrong and that you do not need them at all. You execute undo (u in normal mode) four times in a row (or press 4u) as you normally would, and then the last four changes are gone. Now you see that you need to make an extra change to the file, and you add this also. Normally, your four undone changes would be gone, but not in Vim. When you reverted the four changes, and added the new changes, you actually added another branch to your undo branch tree. In one branch you have the four changes you reverted, and in the other branch you have your most recent change. You could go on like this and add other branches to your undo branch tree. At this point, it might be nice to get an overview of the branches your currently have. This is done with the following command: :undolist
[ 97 ]
Production Boosters
This will get you a list that shows you three pieces of information about each branch—the change number (used to identify a branch), number of changes in a branch, and the time of the branch creation. It could look like this: number changes 6 5 11 8
time ~ 12:12:11 14:01:15
If you want to go to one of the specific change numbers, then simply use: :undo N
where N is the change number. You could also move backwards in the list of changes using the following normal mode command: g- (use g+ if you want to move forward instead).
So what is the difference between using g- and u to go back through the changes? Let’s visualize it with an example. Write the following text in Vim: My name is Jim
Then go to the J and press x three times to delete the name Jim. You now have: My My My My
name name name name
is Jim is im is m is
Now you realize that your name is in fact Jimmy, so you undo the change: My name is m My name is im My name is Jim
You now have one branch with the deletion of the name Jim. Now change the name to Jimmy: My name is Jimm My name is Jimmy
[ 98 ]
Chapter 4
But wait! Your name is actually Kim and not Jimmy. Jim was close to Kim, so let’s undo back to this place using u and change the J to K My My My My
name name name name
is is is is
Jimm Jim im Kim
Another branch was added for when you were undoing the change from Jim to Jimmy. Now let’s go back through the changes with multiple executions of gMy name is Kim My name is im
(Vim changes to new branch) My name is Jim My name is Jimm My name is Jimmy
(Vim changes to new branch) My My My My
name name name name
is is m is im is Jim
Now let’s compare this to using u for undoing the changes: My My My My My My
name name name name name name
is is is is is is
Kim im Jim Jimm Jimmy Jim
As you can see, the u command only takes you directly through the changes that are not in branches while g- takes you through every single change in every branch. So basically, undo branches can give you access to any text state your file content has had.
[ 99 ]
Production Boosters
Instead of going step by step through the changes in all the branches, you could instead jump a time slice back in your 'edit time.' For this Vim has two commands that jump different time slices, back and forth in the undo history depending on the argument. The commands are like this: :earlier Ns :earlier Nm :earlier Nh :later Ns :later Nm :later Nh
Where N is a number of seconds (s), minutes (m) or hours (h) you want to jump back/forward in time. If you use the :undolist function, then you can see the time of the changes and from this you can calculate how far back/forward to jump. It might take some time to get used to having the undo branches, but when you have gotten used to them, they will help you a lot in your work.
Folding
Often when you work with large files, especially code, it can be hard to get a good overview. In Vim there is a special feature that helps you get around this—folding text blocks into folds. In this recipe we will look at how to use folds to make your code easier to overview. A fold is a way of 'folding' a range of lines (e.g. a function scope) into one single line without losing the contents. An example could be the following code: function myFunction(){ var a = 1; var b = 0; var c = a+b; return c; }
If this is folded, then it could be set to look like: +-- 6 lines: function myFunction(){ ------------------------------
In this case the folding follows the syntax of the code and uses the { } to figure out where to do the folding. Besides using syntax, Vim can also do folding according to the following: •
Manual fold marks—manually mark fold (see :help fold-manual)
•
Indent folds—use indentation as fold indication (see :help fold-indent) [ 100 ]
Chapter 4
•
Expression folds—use an expression to find folds (see :help fold-expr)
•
Syntax folds—use syntax as fold indication (see :help fold-syntax)
•
Diff folds—fold unchanged text (see :help fold-diff)
•
Marker folds—insert markers in text as fold indication (see :help fold-marker)
Which type of fold indication to use depends on the type of the text you are working on, and also what you find to be the best for you. So let’s look at how you actually do the folding. The first thing to do is to actually activate the functionality: :set foldenable
Now Vim knows that it should watch out for folding commands when in normal mode. There is a range of commands you can use to open and close folds, but the primary ones are: zc : close a fold zo : open a fold zM : close all folds zR : open all folds
If we take the Syntax folding method as an example, then you just need to place the cursor somewhere in the area you want to fold (e.g. inside a function scope) and then you go into normal mode and press zc to close the fold. Now you will see the function get folded into a single line. In the following figure you can see both folded and unfolded code mixed together:
[ 101 ]
Production Boosters
If you don't want to remember commands for both opening and closing folds, then simply bind a key to toggle a fold open or closed. This could be for example, space: :nnoremap <space> za.
If you find that the design of the folded line is not giving you the information you need, then you can change it easily. Simply change the value of the foldtext option to point to another function that returns the line you want: :set foldtext=MyFoldFunction()
The function could look like the following: function! MyFoldFunction() let line = getline(v:foldstart) “ cleanup unwanted things in first line let sub = substitute(line, '/\*\|\*/\|^\s+', '', 'g') “ calculate lines in folded text let lines = v:foldend - v:foldstart + 1 return v:folddashes.sub.'...'.lines.' Lines...'.getline(v:foldend) endfunction
This function changes your folded line to look like: +--function myFunction(){...6 Lines...}----------------------------
You can see that the function used three different variables whose names start with v:. These are variables set by Vim and contain: •
v:foldstart
Line number of first line in fold
•
v:foldend
Line number of last line in fold
•
v:folddashes
Contains a dash (-) for each level of folding a fold is
The last variable gives a fast indication of how many levels down you are in the folding tree. If you have: if (x != y){ if (y !=x){ print “x not y“; } }
Then v:folddashes in the innermost if will contain '--' (second level) but in the outermost if it will contain '-' (first level).
[ 102 ]
Chapter 4
The dashes at the end of the fold line are automatically added. If you want a different character instead of a dash, then this can of course also be changed. If you, for example, want to change it to equal signs (=) instead of dashes then simply do: :set fillchars=fold:=
You might think why dashes?, but there is actually quite an obvious explanation. Vim has another fold setting called foldcolumn that tells it how many columns to the left of the text it should use for fold information. What it actually uses the columns for is to draw an ASCII fold-tree where the dashes in the folds are the leaves. For example: | some text +- a first level fold ---------------------------------| − beginning of open fold. 2 indication of fold level 2 - do - open fold beginning level 1 +-- a second level fold. | more text | more text
As you can see, it basically draws a tree which looks like: | +| +-| ++--
To set how wide the tree should be, simply use: :set foldcolumn=N
where N is a number between 0 and 12. A value of 1 or 2 is recommended only if you have a few levels of folds, else you should use 3-5. You can execute a command on all folds that are either open or closed: :folddoopen cmd – execute cmd on all line not in a closed fold. :folddoclose cmd – execute cmd on all lines in closed folds.
[ 103 ]
Production Boosters
Simple Text File Outlining
Sometimes when you write a simple text file in Vim and you suddenly realize that it has grown from simple to long and chaotic, and you could really use a good way to do outlining of the text. In this recipe, we look at how to use folding in Vim to do just that, especially if you think quite a bit about how you structure your text. Let's say you have a text like: Chapter 1 Section 1 – Vim help here is some text about the vim help system. Section 2 – vim scripts this section contains info about vim scripts.
Now you would like to fold the text such that only the section headers are shown. If you use manual folding (:set foldmethod=manual) this is quite simple. You simply need to mark all lines in a section (including header line) and then press zf. Now you have created a fold containing those lines. If you start from the outside and go inwards, then you can have say 'Chapter 1' as first level, and each section as fold level 2. If you then closed the section folds, it would look like: -Chapter 1 +Section 1 – Vim help (2) +Section 2 – vim scripts (4)
To make it look exactly as the above example, you will need the following settings: :set foldcolumn=1 :set fillchars=fold:\ “there is a space after the \ :set foldtext=getline(v:foldstart).' ('.v:foldstart.')'
As you can see, this looks a lot like the Table of Contents in a book, but the difference is that this is just a simple text file. As long as you add new text to the sections by appending it to the previous lines (newlines are OK), Vim will still know that your added text is part of the fold. If you later want to delete a fold, then you simply mark the text again in visual mode and then press zd. If you want to use a different formatting of your text (e.g. having '= =' around the section headers), you can do that—as long as you mark your own folding areas, then you won't feel a difference. [ 104 ]
Chapter 4
���������������������������������� Using vimdiff to Track the Changes
Sometimes you have multiple versions of the same file—maybe they are the same and maybe they are not. On Unix systems, there has been a program called diff available for many years (first release in 1974), but on other operating systems you most likely do not have this. This program gives the user an output that shows the differences between two files. Vim has a solution for you that gives this functionality and even presents it in an easy-to-overview format, vimdiff. This recipe will show you how to use Vim to get an overview of changes to your file compared to other versions of the same file. Vimdiff is actually a built-in diff program, which uses colors to show the differences between two files (shown in two windows in Vim split vertically or horizontally). The following figure shows how a vimdiff session could look like.
There are several different ways to activate vimdiff. On many systems, a program shortcut is made that is called vimdiff. In those cases, you can simply use: vimdiff file1 file2
This is actually a shortcut for writing vim -d file1 file2
You will need to supply at least two versions of the file, but up to four versions of the file are supported.
[ 105 ]
Production Boosters
If you are already in an open Vim session, then you can of course also activate the diff mode in Vim. To do so, you will have to use one of the following commands: •
:diffsplit filename
•
:vert diffsplit filename
•
:diffthis
Split the window horizontally and show the file “filename“ in both of the windows. All diff-specific settings are set for both windows. Split the window vertically and show the file 'filename' in both of the windows. All diff-specific settings are set for both windows. Add the current window to the existing set of diff-enabled windows. This could be used if you want to diff a file again with yet another version.
An example could be to see which changes you have made to the currently active file since the last time Vim saved a backup of the file (those copies of the files saved with ~ after the name). If you are working on a file called main.c, then you could at any time save the file and then execute the following command: :vert diffsplit main.c~
Now the current window is split vertically into two, and you will be able to see the changes you have made to the file marked with colors (depending on your color theme). Maybe it wasn't actually you who made the changes to one of the versions of the file and maybe another developer just sent you a patch for your file. But don't worry, Vim can still give you a nice diff view of the changes the patch makes to your version of the file. Simply open your version of the file and then with it in the active window, execute this command: :vert diffpatch patchfile
(or just :diffpatch patchfile)
where patchfile is the patch the other developer sent you. Now Vim will open another window with your file in, and then apply the patch to it. Then it will set up all the diff settings for the windows, such that the changes are colored.
Navigation in vimdiff
Navigation in the vimdiff windows is a bit different than in the normal Vim windows. When for instance you scroll through the file in one of the windows in the diff split, you will see that the other part of the diff windows is also scrolled. In fact, the windows follow each other in a way such that the current line in one window is [ 106 ]
Chapter 4
the same line (if available) as in the other window. This is called scrollbind and can be turned on and off with: :set scrollbind
and: :set noscrollbind
You can edit the files in the diff windows and you will see that the diff colors are updated accordingly; if not, then try executing: :diffupdate
When you are in one of the diff windows, you might want to jump fast between changes done to the file. This is done with the following commands in normal mode: [c
: Go to start previous change
]c
: Go to start of next change
This way you can navigate between relevant areas in a file and get a good idea of which changes have been made to the file. When the cursor is placed in one of the changes you can see in the vimdiff window, then you might realize that this change is also needed in the other version of the file. So now you could copy the lines in the change, and insert them into the other file in the right place. But why use time on this, when Vim has made it a lot easier for you? Vim has a function that simply puts the change from one file version into the correct place in another version of the same file. The command is: :diffput
This command should be executed when the cursor is placed in the change you want to move to the other file version. If on the other hand you are in the file without the change, then you could either move to the other file and then put the change back into the first file—or use the Vim way of doing it, with the following command: :diffget
Alternatively, you can use do in normal mode for getting a change and dp for putting a change. Read more about vimdiff in :help vimdiff
[ 107 ]
Production Boosters
Using Diff to Track Changes
From the previous section, we know how to use vimdiff to make a diff of different versions of the same file. But what if you really just want to know what you have changed in the current buffer before saving it? This recipe gives you a little trick to check for changes between the version of the file you have on your hard drive and the one you have in the buffer. It will, in other words, show you what you have changed in the buffer since the last time you saved. What you have to do is to add the following function to your vimrc file:
function! DiffWithFileFromDisk() let filename=expand('%') let diffname = filename.'.fileFromBuffer' exec 'saveas! '.diffname diffthis vsplit exec 'edit '.filename diffthis endfunction
The function stores a temporary copy of the file you have in the current buffer (including latest changes) and then it diffs this file against the version of the file you have on your hard drive. To call the function you use the following command: :call DiffWithFileFromDisk()
or if you want to access it faster, then you can bind a key to the command like this: :nmap :call DiffWithFileFromDisk()
This binds the F7 key to the function call, and you then just have to go into normal mode and press F7 to see the changes marked in diff mode. Now you can quickly and easily go through the changes that you have made to the file and check if all of them are important.
Open Files Anywhere
System administrators and web developers have one thing in common, which can be a big point of annoyance for both of them, if they don't have an easy way to get around it. They both work with files that are most often placed on remote servers of some sort. [ 108 ]
Chapter 4
The system administrator mostly gets around the problem by logging in on the remote server via for example an SSH (secure shell) connection and then edits the configuration files etc. directly on the server. The web developer on the other hand gets around the problem by uploading and downloading the files between the remote computer and his or her local computer using an FTP client or by using systems like Webdav. But what if it didn't have to be that way? What if they could simply edit the files on the remote system directly from their local system? In Vim, this is in fact possible, without any further extensions besides what normally comes with it. Vim has a system called netrw (for net read/write), which comes in handy whenever you want to work with files on remote servers. Let's get right to it and start with an example. Imagine a web developer, John, who has his homepage placed on a remote system called remote.server.com. He wants to edit his index.html file, which resides in the public_html/ directory in his home directory on the server. In this case, the web developer could simply open the file in Vim like this: vim ftp://
[email protected]/public_html/index.html
Vim recognizes that it needs to use the FTP protocol and then connects to the FTP server on remote.server.com using john as username. If a password is needed to log in, then Vim will prompt you for it. Vim transfers a temporary copy of the file to the local machine and then lets you edit it like any other file. Only difference is that whenever you save the file, it is saved onto the remote server also. If he had already opened Vim, then John could instead open the site with one of the following commands: :Nread ftp://
[email protected]/public_html/index.html :Nread remote.server.com john PASSWORD public_html/index.html
Change PASSWORD to be the password you want to use for the FTP server. Besides reading a file from a remote server, you can also open a local file and write it to the remote server—or go about opening a file on one remote server and save it onto another. The command for writing a file to a remote FTP server is: :Nwrite ftp://user@server/path/to/filename :Nwrite server user password path/filename
The format of the arguments from :Nread and :Nwrite can differ from protocol to protocol. Use :Nread ? and :Nwrite ? to get help on the exact syntax.
[ 109 ]
Production Boosters
Besides the FTP protocol, Vim supports many other protocols: •
SCP
•
SFTP
•
RCP
•
HTTP (read-only)
•
DAV
•
rsync (read-only)
•
fetch (read-only)
To use these, you simply change the 'ftp' part of the previous example to one of the other protocol names (in lowercase). There is, however, a catch. Vim is dependent on external command-line programs in order to use the different protocols. On Linux systems most of these programs are available by default, but on Mircosoft Windows, only FTP is available. You can find a list of the external programs that Vim uses as a default and explanation of how to change them in the help system: :help netrw-externapp
Besides reading and writing files, Vim is also able to give you a directory listing, such that you can use it for finding the right remote files to edit. You just have to point your :Nread to a directory instead of a file. For example: :Nread scp://user@server/some/directory/
You can select any file in the directory listing and it will then be opened in Vim as if it was a local file. If you use Linux, then you can store usernames and passwords for remote sites in a .netrc file in your home directory. See :help netrw-netrc for more information.
Faster Remote File Editing
So now you have learned how to work with remote files directly, and you will at some point get in a situation where you have several remote files open at the same time. But then you suddenly hit an annoying situation—you need to re-log in every time you move to another buffer with for example :bufferprev and :buffernext.
[ 110 ]
Chapter 4
As a default, Vim tries to reload the contents of a buffer whenever it is shown in a window. This means that if the file in the buffer is a remote file, then Vim will need to log in again, in order to check if the file should be reloaded. But is that actually necessary? If you can live with the fact that a remote file can be edited by another person while you are editing it remotely (without you being notified about it), then you can trick Vim into not reloading the file. Each buffer has a set of options that tells Vim what to do with the particular buffer in different situations. One of the options is bufhidden, which tells Vim what to do when a buffer is hidden (not shown in a window). This option is normally not set to anything, but if you set it to 'hide', then you tell Vim to just hide the buffer when it is not in a window, and then just show it again when you show it again in a window. Simply, add the following to your vimrc file: :set bufhidden=hide
And that's basically it. Now you won’t have to re-log in whenever you switch buffer and will feel just as if you were editing a local file.
Summary
In this chapter we have been looking at how to improve our daily work in Vim. Many approaches have been touched, each with a specific area to optimize. We started out by looking at how to use templates to minimize the amount of text to enter. The first time, our templates were simple ones that used the abbreviation functionality in Vim to emulate the insertion of a template. Next, we improved on our template system by creating template files for specific file types and thereby made it possible to insert an entire skeleton into a programming file. After the templates, we moved on and looked at another way to minimize word entry time—auto-completion. Different approaches for auto-completion were discussed and a function was proposed for binding all auto-completion to a single key—Tab. By recording a list of commands, you can get around the boring task of entering the same commands over and over again. We looked at how to do Vim macro recording and use this to change a simple text file into a fine HTML file in a matter of minutes. Next up, was Vim Sessions and how to use these for everything from saving the look of a window, to using sessions as a full-fledged project manager.
[ 111 ]
Production Boosters
With registers, you have the possibility to use not one, not two, but nine different registers/clipboards. Using folding of the text, you can get a better overview of a file because every unnecessary part is hidden in a fold. You can even use folding for creating simple outlines of a text file. So now the file is changed, but what has actually been touched in the file? We looked at how to use the built-in diff functionality in Vim. This gives you an improved overview of where and what you have changed in the active files and you can even undo or add new changes while still maintaining the good diff overview. Working with files on the local machine is one thing, and working on remote files, another. In Vim, it is possible to work in and navigate remote files directly. This way you won't feel whether the file is local or remote. After reading this chapter and playing around with the recipes, you should soon feel how your workday has improved.
[ 112 ]
Advanced Formatting Often the simplest modification to a text or a piece of code is what changes it from being obscure to being easily readable. In this chapter, we will look at some of the simple tricks you can use to format the text you are working on—no matter whether it is plain text or code. This chapter will have recipes in three categories: •
Text formatting
•
Code formatting
•
Using external formatter programs
After reading this chapter, you should have a good idea about what is possible and what is not, when it comes to text formatting in Vim.
Formatting Text
Even though most people prefer graphical word processors like Microsoft Word or OpenOffice Writer when they want to write plain text, there are still times where an editor like Vim will do it just as well. In the following sections, we will look at how to use the strengths of Vim when formatting normal text.
Putting Text into Paragraphs
This recipe is probably one of the simplest in this book, but at the same time one of the most versatile when it comes to formatting plain text. Imagine that you are writing a piece of text and just keep on writing without bothering about changing lines or formatting the text. At some point, you might end up with one or more very long lines and conclude that you should start formatting the text. At this point, you have two choices:
Advanced Formatting
1. Go through the text and add the formatting manually 2. Use the strength of Vim and format the entire paragraph with one command Obviously, the latter option is the fastest one and at the same time the formatting will be consistent. So let’s look at which command to use for this: gqap
This command is actually a combination of a command and a movement; specifically: gq
: format everything the next movement moves over
ap
: 'a paragraph' moves over the current paragraph
In other words, the command simply tells Vim to move over the current paragraph and format it. A paragraph is defined as all lines between two empty lines. In order to change to another paragraph, you simply add an empty line. The formatting that Vim adds to the text is basically nice line breaks such that the lines are not longer than a specific length (split correctly between words). The text formatting width is defined in the Vim option textwidth such that, if you want a maximum of 80 characters on each line, then you would need to have the following in your vimrc file: :set textwidth=80
If the option is set to 0, then Vim sets it to the width of the window—however, never more than the number of characters defined in the textwidth setting. How Vim formats a paragraph can be set in the Vim option called formatoptions. See :help 'formatoptions'and:help 'fo-table'. gq can be used together with any movement command, and after performing the
formatting, it will place the cursor where it ends (typically at the end of the last line in that particular area). If instead you want the cursor to go back to the place where it originally was before executing the command, then simply change gq to gw. If you have the cursor placed at the beginning of the first line and do gwap, then the cursor will remain there even though the paragraph is formatted. You can repeat the formatting multiple times by prepending the command with the number of times it should be repeated, for example 5gqap will format the current and the next four paragraphs. If you want to format all the paragraphs in a file, you can do it with the command 1gqG. [ 114 ]
Chapter 5
This formatting command does not only apply to plain text, but also to any other type of content—and you can decide what formatting it should apply. You can set any function to be the 'formatter' for any given file format, simply by setting it in the Vim option called formatexpr. If for instance you work with a C-code source file, then you should simply have the following in your vimrc file: :set formatexpr=c#Formatter()
This tells Vim that when it opens a file of type c, it should use the function called Formatter() in the autoloaded file for the C filetype. Autoloaded files can be found in your VIMHOME in a folder called autoload. Files are named as the file type and appended with .vim. For example, VIMHOME/autoload/c.vim for the C filetype.
A formatting function has three variables that you can use to find the text you have told it to format: •
v:num
•
v:count The number of lines to format.
•
v:char
The line number of the first line to format. This variable holds a character that is going to be inserted. This can be empty.
A simple formatting function could look like this: function! MyFormatter() let first = v:num let last = v:num + v:count while(first. This basically means from first selected line ('") let difsize = ending-beginning +1 let pre = ' ' while (beginning 2,{3,}3”, which means that we want Vim to add two extra spaces to the normal indent length, and we want to place {and} three spaces in compared to the previous line. So, if we have normal indent to be four spaces, then the above example could result in code looking like this (dot marks a space):
if( a == b) ...{ ......print “hello”; ...}
The default value for cinoptions is this quite long string: “>s,e0,n0,f0,{0,}0,^0,:s,=s,l0,b0,gs,hs,ps,ts,is,+s,c3, C0,/0,(2s,us,U0,w0, W0,m0,j0,)20,*30” . See :help 'cinoptions' for more information on all
the options.
[ 122 ]
Chapter 5 •
cinwords
This option contains all the special keywords that will make Vim add indentation on the next line. An example could be: :set cinwords=”if,,else,do,while,for,switch”, which is also the default value for this option. See :help 'cinwords' for more information.
Indentexpr
Indentexpr is the most flexible indent option to use, but also the most complex. When used, indentexpr evaluates an expression to compute the indent of a line— hence you have to write an expression that Vim can evaluate. You can activate this option by simply setting it to a specific expression: :set indentexpr=MyIndenter()
Where MyIndenter() is a function that computes the indentation for the lines it is executed on. A very simple example could be a function that emulates the autoindent option: function! MyIndenter() " Find previous line and get its indentation let prev_lineno = s:prevnonblank(v:lnum) let ind = indent( prev_lineno ) return ind endfunction
Adding just a bit more functionality than this, the complexity increases quite fast. Vim comes with a lot of different indent expressions for many programming languages. These can serve as inspiration if you want to write your own indent expression. You can find them in the folder indent in your VIMHOME. You can read more about how to use indentexpr in :help 'indentexpr' and :help 'indent-expression'.
Fast Code Block Formatting
After you have configured your code formatting, you might want to update your code to follow these settings. To do so, you simply have to tell Vim that it should reindent every single line in the file from first line to the last. This can be done with the following Vim command: 1G=G
[ 123 ]
Advanced Formatting
If we split it up it simply says: 1G = G
: Go to first line of file (alternatively you can use gg). : Equalize lines, in other words indent according to formatting configuration. : Go to the last line in file (tells Vim where to end indenting).
You could easily map this command to a key in order to make it easily accessible: :nmap 1G=G :imap <ESC>1G=Ga
The last a is to get back into insert mode as this was where we originally were. So now you can just press the F11 key in order to reindent the entire buffer correctly. Note that if you have a programmatic error, like for example missing a semicolon at the end of a line in a C program, the file will not be correctly indented from that point on in the buffer. This can sometimes be useful to identify where a scope is not closed correctly (for example a {not closed with a } ).
Sometimes you might just want to format smaller blocks of code. In those cases, you typically have two options—use the natural scope blocks in the code, or select a block of code in visual mode and indent it. The last one is simple. Go into visual mode with e.g. Shift-v and then press = to reindent the lines. When it comes to using code blocks on the other hand, there are several different ways to do it. In Vim there are multiple ways to select a block of code, so in order to combine a command that indents a code block, we need to look at the different types and the commands to select them: •
i{
'Inner block', which means everything between { and } excluding the brackets. This can also be selected with i} and iB.
•
a{
'A block', which means all the code between { and } including the brackets. This can also be selected with a} and aB.
•
i(
'Inner parenthesis', meaning everything between ( and ) excluding the parentheses. Can also be selected with i) and ib.
•
a(
'A parentheses, meaning everything between ( and ) including the parenthesis'. Can also be selected with a) and ab.
•
i<
'Inner block', meaning everything between < and > excluding the brackets. Can also be selected with i>.
[ 124 ]
Chapter 5 •
a<
'A block', meaning everything between < and > including the brackets. Can also be selected with a>.
•
i[
'Inner [] block', meaning everything between [ and ] excluding the square brackets. Can also be selected with i].
•
a[
'A [] block', meaning everything between [ and ], including the square brackets. This can also be selected with a].
So we have defined what Vim sees a block of code as; now we simply have to tell it what to do with the block. In our case, we want to re-indent the code. We already know that = can do this, so an example of a code block re-indentation could look like: =i{
Which, if executed at the following code (| being the place where the cursor is): if( a == b ) { print |“a equals b”; }
would produce the following code (with default C format settings): if( a == b ) { print |“a equals b”; }
If on the other hand we choose to use a{ as the block we are working on, then the resulting code would look like: if( a == b ) { print “a equals b”; }
As you can see in the last piece of code, the =a{ command corrects the indentation of both the brackets and the print line. In some cases, where you work in a code block with multiple levels of code blocks, you might want to re-indent the current block and maybe the surrounding one. No worries, Vim has a fast way to do this. If for instance you want to re-indent the current code block, and besides that want to re-indent the block that surrounds it, you simply have to execute the following command while the cursor is placed in the innermost block: =2i{ [ 125 ]
Advanced Formatting
This simply tells Vim that you will equalize/re-indent two levels of inner blocks counting from the “active” block and out. You can replace the number '2' with any number of levels of code blocks you want to re-indent. Of course, you can also swap the inner block command with any of the other block commands, and that way select exactly what you want to re-indent. So, this is really all it takes to get your code to indent according to the setup you have.
Auto Format Pasted Code
The trend among programmers tells us that we tend to reuse parts of our code—so called patterns. This could mean that you have to do a lot of copying and pasting of code. Most users of Vim have experienced what is often referred to as the 'stair effect' when pasting code into a file. This effect occurs when Vim tries to indent the code as it inserts it. This often results in each new line to be indented another level, and you ending up with 'a stair': code line 1 code line 2 codeline 3 code line 4 ...
The normal workaround for this is to go into paste-mode in Vim, which is done by using: :set paste
After pasting your code, you can now go back to your normal insert mode again: :set nopaste
But what if there was another workaround? What if Vim could automatically indent the pasted code such that it is indented according to the rest of the code in the file? Vim can do that for you with a simple paste command. p=`]
This command simply combines the normal paste command (p) with a command that indents the previously inserted lines (=`]). It actually relies on the fact that when you paste with p (lowercase), then the cursor stays on the first character of the pasted text. This is combined with `], which takes you to the last character of the latest inserted text, and gives you a motion across the pasted text from first line to the last. [ 126 ]
Chapter 5
So all you have to do now is to map this command to a key and then use this key whenever you paste a piece of code into your file. If you want to use the normal p pasting key but with the new functionality, then you can use the following mapping: :nnoremap p p=`] :nnoremap p
It maps Ctrl-p to what p normally did (past with no formatting) and then maps p to the new paste with automatic formatting.
Using External Formatting Tools
Even though experienced Vim users often say that Vim can do everything, this is of course not the truth, but is close. For those things that Vim can’t do, it is smart enough to be able to use external tools. In the following sections, we will take a look at some of the most used external tools that can be used for formatting your code, and how to use them.
Indent
The Indent program is probably one of the most used external programs for Vim. It has been around since the late 80s for various Unix platforms, and has also later been imported to other platforms including Microsoft Windows. As the name indicates, this program indents code—especially code that resembles C code in syntax. What you may wonder is why you would use an external program for this, when Vim can handle this task just fine. This is a good question because Vim can do this very well, but the Indent program does it better—and at the same time making it easier to standardize the indentation among multiple editors . By specializing only in indenting code, indent is able to more effectively indent code than the limited indent functionality included in Vim, for which indenting is a feature and not 'the feature'. Indent specializes in understanding the code, and indents it according to the code—even if there is a syntactic error in the code. So how do you use Indent from within Vim? Previously, we have seen several different options for how Vim should indent your code. There is, however, one option that overrules all of them: :set equalprog=PROGRAM [ 127 ]
Advanced Formatting
What this option does is set the external program Vim should use for indentation when using the commands with = . In the case of Indent, you simply change PROGRAM to the path to your Indent program. Now, whenever you use one of the indentation commands like 1G=G, it takes the involved lines and pipes them through the program you have defined in equalprog. You can even supply the program with command-line arguments, if needed. In the case of Indent, there are so many different command line arguments that you will get a better result by configuring its configuration file. You can always find the latest version of the Indent program at this address: http://mysite.wanadoo-members.co.uk/indent/ beautify.html.
Berkeley Par
In the early 90s, Adam M. Costello began working on a simple command-line program whose only purpose was that when given a text with a paragraph in it, the program would reformat it according to the user’s wishes. The program was called Par and, within a year or two, it evolved into a very feature-rich program that can re-format nearly any type of paragraph. This of course makes Par an ideal external friend for Vim, so let’s look at some examples of how it can be used. If, for instance, you want your text to be nicely formatted in paragraphs with no more than 78 characters on each line, then you could simply use it as: :set formatprg=par\ -w78
The formatprg option in Vim tells it which program to use for formatting of the text when one of the gq commands is used. Notice that the space between the program name and its option is escaped with a backslash. This is needed in order for Vim to see the entire string as one option and not two. Note that Vim will only use formatprg when formatexpr is empty. Otherwise the formatexpr will be used.
From earlier on, we know that Vim cannot justify the text such that both ends of the lines are aligned with the margins. Fortunately, Par can help us here, and simply by adding a j (for 'justify') to our previous formatprg value, we can get Par to justify the text: :set formatprg=par\ -w78j [ 128 ]
Chapter 5
Par can not only be used on normal text, but also some parts of the code— the comments. If you, for instance, have the following comment: /********************************************************************/ /* This function helps you modify a string and remove all */ /* unnecessary characters . */ /* Don't use this on widechar strings or strings shorter than 10 */ /* characters */ /********************************************************************/
You could select it in Vim and then do: !par 60r
(Vim adds ' MyscriptMyfunctionA <SID>MyfunctionA noremap <SID>MyfunctionA :call <SID>MyfunctionA()
The first mapping maps our unique MyscriptMyfunctionA identifier to <SID>MyfunctionA. We use <SID> here, because this little tag is exchanged with Vim's own unique ID for the current script, and this is needed if we want to make a global mapping to a function that is only available in script scope (e.g. s:MyfunctionA). The second mapping binds the actual function (<SID>MyfunctionA(), which is s:MyfunctionA()) to the global mapping <SID>MyfunctionA. So what actually happens is that when you press \a (having mapleader set to \), then your first mapping translates this into MyscriptMyfunctionA. This is defined in the script and hence the <SID> is now has the right value. Therefore Myscr iptMyfunctionA is again translated further into <SID>MyfunctionA which is finally mapped into the actual call of the local function s:MyfunctionA().
[ 167 ]
Vim Scripting
You might find this complicated and a bit too much, and you might be right with a relatively unique function name like MyfunctionA(). But what if the function was called Add(), Delete(), Convert() or some other function name that many scripts could have implemented. In those cases the function names would clash and Vim would not know which one to use. You could of course just give your functions some weird unique names, but that will in the end just make your script code cluttered and pollute global scope with unnecessary functions. For more info see: ����������� :help <SID> ������������ :help :help 'script-local'
Functions It is now time to add the functions to the script file. We have already seen how a function is added and noticed that it might be better that all functions that are not directly needed in global scope are put into the script scope with the s: scope marker. So an example of this could be: " this is our local function with a mapping function s:MyfunctionA() echo "this is the script-scope function MyfunctionA speaking" endfunction " this is a global function which can be called by anyone function MyglobalfunctionB() echo "Hello from the global-scope function myglobalfunctionB" endfunction " this is another global function which can be called by anyone function MyglobalfunctionC() echo "Hello from MyglobalfuncionC() now calling locally:" call <SID>MyfunctionA() endfunction
The first function is a private function, which is only available in the script scope, while the two others are both available in the global scope. Notice, however, how it is possible for one of the global functions to call the local function because it knows the correct <SID> for the current script. Putting it All Together So let's put it all together and see it as a full script example. " " " "
myscript.vim Version Maintainer Last modified
– : : :
Example script to show how a script is structured. 1.0.5 Kim Schulz 01/01/2007 [ 168 ]
Chapter 6 " License
: This script is released under the Vim License.
" check if script is already loaded if exists("loaded_myscript") finish "stop loading the script endif let loaded_myscript=1
let s:global_cpo = &cpo "store compatible-mode in local variable set cpo&vim " go into nocompatible-mode " ######## CONFIGURATION ######## " variable myscript_path if !exists("myscript_path") let s:vimhomepath = split(&runtimepath,',') let s:myscript_path = s:vimhomepath[0]."/plugin/myscript.vim" else let s:myscript_path = myscript_path unlet myscript_path endif " variable myscript_indent if !exists("myscript_indent") let s:myscript_indent = 4 else let s:myscript_indent = myscript_indent unlet myscript_indent endif
" ######## FUNCTIONS ######### " this is our local function with a mapping function s:MyfunctionA() echo "This is the script-scope function MyfunctionA speaking" endfunction " this is a global function which can be called by anyone function MyglobalfunctionB() echo "Hello from the global-scope function myglobalfunctionB" endfunction " this is another global function which can be called by anyone function MyglobalfunctionC() echo "Hello from MyglobalfuncionC() now calling locally:" call <SID>MyfunctionA()
[ 169 ]
Vim Scripting endfunction " return to the users own compatible-mode setting :let &cpo = s:global_cpo
And there you have it! Our very first Vim plugin script! It might not have that much functionality, but it shows very well how you should structure your script to make it more understandable. Vim has other types of scripts like file-type plugins, compiler plugins, and library scripts. You can read more about how to modify your script in order to make it these types of scripts in: :help 'write-filetype-plugin' :help 'write-compiler-plugin' :help 'write-library-script'
On the Vim online community site, you will find thousands of scripts, which you can use as inspiration. Some of these are even library scripts that add functions you can use in your own script to speed up development. See : http://www.vim.org.
Scripting Tips
In this section, we will look at a few extra tips that can be handy when you create scripts for Vim. Some are simple code pieces you can add directly in your script, while others are good-to-know tips.
Gvim or Vim?
Some scripts have extra features when used in the GUI version of Vim (Gvim). This could be adding menus, toolbars, or other things that only work if you are using Gvim. So what do you do to check if the user runs the script in a console Vim or in Gvim? Vim has already prepared the information for you. You simply have to check if the feature gui_running is enabled. To do so, you use a function called has(), which returns 1 (true) if a given feature is supported/enabled and 0 (false), if not. An example could be: if has("gui_running") "execute gui-only commands here. endif
This is all you need to do to check if a user has used Gvim or not. Note that it is not enough to check if the feature "gui" exists, because this will return true if your Vim is just compiled with GUI support—even if it is not using it. [ 170 ]
Chapter 6
Look in :help 'feature-list' for a complete list of other features you can check with the has() function.
Which Operating System?
If you have tried to work with multiple operating systems like Microsoft Windows and Linux, you will surely know that there are many differences. This can be everything from where programs are placed, to which programs you have available and how access to files is restricted. Sometimes, this can also have an influence on how you construct your Vim script as you might have to call external programs, or use other functionality, specific for a certain operating system. Vim lets you check which operation system you are on, such that you can stop executing your script or make decisions about how to configure your script. This is done with the following piece of code: if has("win16") || has("win32") || has("win64")|| has("win95") " do windows things here elseif has("unix") " do linux/unix things here endif
This example only shows how to check for Windows (all flavors available) and Linux/Unix. As Vim is available on a wide variety of platforms, you can of course also check for these. Find all the operating systems in: :help 'feature-list'
Which Version of Vim?
Throughout the last decade or two, Vim has developed and been extended with a large list of functions. Sometimes, you want to use the newest functions in your script, as these are the best/easiest to use. But what about the users who have a version of Vim that is older than the one you use, and hence don't have access to the functions you use?
[ 171 ]
Vim Scripting
You have three options: 1. Don't care and let it be the user's own problem (not a good option). 2. Check if the user uses a old version of Vim, and then stop executing the script if this is the case. 3. Check if the user has too old a version of Vim, and then use alternative code. The first option is really not one I would recommend anyone to use, so please don't use that option. The second option is acceptable, if you can't work around the problem in the old version of Vim. However, if it is possible to make an alternative solution for the older version of Vim, then this will be the most preferable option. So let's look at how you can check the version of Vim. Before we look at how to check the version, we have to take a look at how the version number is built. The number consists of three parts: •
Major number (e.g. 7 for Vim version 7)
•
Minor number (e.g. 3 for Vim version 6.3)
•
Patch number (e.g. 123 for Vim 7.0.123)
The first two numbers are the actual version number, but when minor features/ patches are applied to a version of Vim, it is mostly only the patch number that is increased. It takes quite a bit of change to get the minor number to increase, and a major part of Vim should change in order to increase the major version number. Therefore, when you want to check which version of Vim the user is using, you do it for two versions—major+minor version and patch number. The code for this could look like: if v:version >= 702 || v:version == 701 && has("patch123") " code here is only done for version 7.1 with patch 123 " and version 7.2 and above endif
The first part of the if condition checks if our version of Vim is version 7.2 (notice that minor version number is padded with 0 if less than 10) or above. If this is not the case, then it checks to see if we have a version 7.1 with patch 123. If patch version 124 or above is included, then you also have patch 123 automatically.
[ 172 ]
Chapter 6
Printing Longer Lines
Vim was originally created for old text terminals where the length of lines was limited to a certain number of characters. Today, this old limitation shows up once in a while. One place where you meet this limitation of line length is when printing longer lines to the screen using the "echo" statement. Even though you use Vim in a window where there are more than the traditional 80 characters per line, Vim will still prompt you to press Enter after echoing lines longer than 80 characters. There is, however, a way to get around this, and make it possible to use the entire window width to echo on. By window width, the total number of columns in the Vim window minus a single character is meant. So if your Vim window is wide enough to have 120 characters on each line, then the window width is 120-1 characters. By adding the following function to your script, you will be able to echo screen-wide long lines in Vim: " WideMsg() prints [long] message up to (&columns-1) length function! WideMsg(msg) let x=&ruler | let y=&showcmd set noruler noshowcmd redraw echo a:msg let &ruler=x | let &showcmd=y endfunction
This function was originally proposed by the Vim script developer Yakov Lerner on the online Vim community http://www.vim.org.
Now whenever you need to echo a long line of text in your script, instead of using the echo statement you simply use the function Widemsg(). An example could be: :call WideMsg("This should be a very loooong line of text")
The length of a single line message is still limited, but now it is limited to the width of the Vim window instead of the traditional 80-1 characters.
Debugging Vim Scripts
Sometimes things in your scripts do not work exactly as you expect them to. In these cases, it is always good to know how to debug your script.
[ 173 ]
Vim Scripting
In this section, we will look at some of the methods you can use to find your error. Well structured code often has fewer bugs and is also easier to debug.
In Vim, there is a special mode for doing script debugging. Depending on what you want to debug, there are some different ways to start this mode. So let's look at some different cases. If Vim just throws some errors (by printing them at the bottom of the Vim window), but you are not really sure where it is or why it happens, then you might want to try to start the Vim directly in debugging mode. This is done on the command line by invoking Vim with the -D argument. vim -D somefile.txt
The debugging mode is started when Vim starts to read the first vimrc file it loads (in most cases the global vimrc file where Vim is installed). We look at what to do when you get into debug mode in a moment. Another case where you might want to get into debug mode is when you already know which function the error (most likely) is in, and hence just want to debug that function. In that case you just open Vim as normal (load the script with the particular function if needed) and then use the following command: :debug call Myfunction()
where everything after the :debug is the functionality you want to debug. In this case, it is a simple call of the function Myfunction(), but it could just as well be any of the following: :debug read somefile.txt :debug nmap ,a :call Myfunction() :debug help :debug
So let's look at what to do when we get into the debugging mode. When reaching the first line that it should debug, Vim breaks the loading and shows something like: Entering Debug mode. Type "cont" to continue. cmd: call MyFunction() >
[ 174 ]
Chapter 6
Now you are in the Vim script debugger and have some choices for what to make Vim do. If you are not familiar with debugging techniques, it might be a good idea to read up on this subject before starting to debug your scripts.
The following commands are available in the debugger (shortcut in parentheses): •
cont (c)
•
quit (q)
•
interrupt (i)
•
step (s)
•
next (n)
•
finish (f)
Continue running the scripts/commands as normal (no debugging) until next breakpoint (more about this later). Quit the debugging process without executing the last lines. Stop the current process like quit, but go back to the debugger. Execute next line of code and come back to the debugger when it is finished. If line calls a function or sources a file, then it will step into the function/file. Execute the next command and come back to the debugger when it is finished. If used on a line with a function call it does not go into the function but steps over it. Continue executing the script without stopping on breakpoints. Go into debug mode when done.
So now you simply execute the different commands to go through the lines of the script/function to see how it jumps through the if conditions, etc. If you want to execute the same command multiple times, you simply press Enter without feeding a new command. You can at any point execute another Ex command if needed (see :help 'ex-command-index'), but note that you don't have direct access to the variables etc., in the debugger, unless they are global. Sometimes, the place you want to get to is many lines into the code, and you really don't want to step all the way through the code until you get to this place. In that case, you can insert a breakpoint at the exact line where you want to start the real debugging, and then just execute a cont as first command. A breakpoint is inserted by one of the following commands, depending on how you want it inserted: breakadd func linenum functionname breakadd file linenum filename breakadd here [ 175 ]
Vim Scripting
•
The first example sets a breakpoint on a particular function. The functionname can be a pattern like Myfunction* if you, for instance, want to break on any function with a name that begins with Myfunction.
•
Sometimes, however, it is not in a function that the problem resides, but rather around a specific line in a file. If this is the case, then you should use the second command, where you give it a line number and a file name pattern as arguments to tell it where to break.
•
The final command is used if you have already stepped to the right place in the file but want to be able to break on it the next time you go through the code in the debugger. This command simply sets a breakpoint on the current line, in the current file, where you currently are in the debugger.
•
You can at any point of time get a list of breakpoints with the following command:
:breaklist
•
If a breakpoint is no longer needed, you have to delete it. As when adding breakpoints, there are also few different ways to delete them.
•
The simplest way to do it is by simply finding the number of the breakpoint in the list of breakpoints, and then using the following command:
:breakdel number
•
Alternatively, you can delete the breakpoints the same way as you added them—except that you now use breakdel instead of breakadd:
:breakdel func linenum functionname
:breakdel file linenum file
breakdel here
•
If you want to remove all breakpoints, you can do it in one step by using this command: :breakdel *
You can add a breakpoint directly on the command line when going into debug mode. Simply use the -c argument like: vim -D -c 'breakadd file 15 */.vimrc' somefile.txt
[ 176 ]
Chapter 6
Distributing Vim Scripts
Now that your script is ready, it is time for you to distribute the script (if you have chosen to do so). The online Vim community has become the de facto place to publish scripts for others to find. Because of this, I urge you to do the same. But before you get to this, there are a couple of things you have to get ready. First of all, you need to figure out whether your script needs to be packed into a compressed file like a ZIP file, or if it should just be distributed as a single .vim file. The main reason for choosing the first option is that your script consists of multiple files (like main script, file type plugin, syntax file, documentation, etc.). How to create ZIP files (or related file types) is beyond what this chapter will look at, but here are a couple of pointers on how I make my ZIP files "install ready": •
Create the ZIP file including the folders where the files are placed relative to your VIMHOME. For example, if you have: VIMHOME/plugin/myscript.vim VIMHOME/syntax/mylang.vim VIMHOME/doc/myscript.txt
then, the ZIP file should contain the three folders: plugin, syntax, and doc with one file in each. This makes the installation easy, as you simply have to go into your VIMHOME and then unpack the ZIP file. •
Always include a help file for your script. This file should be installed in VIMHOME/doc/ and contain descriptions of what the script does, which settings it has, and how to use it.
Even though you only have one script file, it can still be a good idea to put it in a ZIP file together with a help file. This makes it easier for you to remember to add documentation. We will look more at how to create Vim documentation in the next section.
Making Vimballs
Another, maybe even more interesting, alternative is to make a Vimball. We have previously looked at how to use Vimballs to install scripts, so it could now be interesting to look a bit at how to create one. The command to create a vimball is constructed as: :[range]MkVimball filename.vba
[ 177 ]
Vim Scripting
This sure seems simple, right? And it really is. There is, however, a bit of preparation you need to do before calling this function. The first thing you have to do is to open a new empty buffer in Vim with: :enew
Now you add the paths to all the files (one on each line) relative to your VIMHOME. To take the above ZIP file example, it could look like: plugin/myscript.vim doc/myscript.txt syntax/mylang.vim
When this is done you are actually ready to execute the command across the range of lines. Place the cursor on the first line in the buffer, go into normal mode, and use Shift-v to select all the lines with paths on. Now all you have to do is to execute the command: :MkVimball myscript.vba
Vim will automatically add the range of the lines you have selected in front of the command. The filename myscript.vba can be any name, but if the file already exists, then a warning will be given, but no file is written. If you really want to overwrite an existing file, then just add an ! after the MkVimball to tell Vim that you mean it. There is not more to it. You now have a vimball file called myscript.vba, which can be installed as described earlier in this chapter. Remember that you need to install the vimball script in order to use the vimball functions we have described here. The latest version of the Vimball script can always be found here: http://www.vim.org/ scripts/script.php?script_id=1502.
Remember the Documentation
Vim has a very comprehensive help system with help for nearly any aspect of using Vim. But what happens when a user installs your script and wants to find help about it? If you haven't added documentation with your script installation, then the user will be out of luck and find no help in the help system. But why not follow the good concept that Vim has started about documenting everything? In other words: "Please include documentation when you release a script for Vim." So let's take a look at how to create Vim documentation so that it has links, markers, tags, etc. just like the real Vim documentation. [ 178 ]
Chapter 6
A Vim documentation document is just a plain text file with some special markup. When you create a new document, the first line is the most important of them all. So let's start by looking at that line: *docname.txt* single line of description
The first * should be the very first character on the very first line of the file. The docname.txt is the name of the file you are currently editing. Vim uses this when linking to the file from the "local additions" list in the Vim help system (see :help local-additions). The description after the second * is a one-line description of what this document is all about. In the case of our example, this line could be: *myscript.txt* Documentation for example script myscript.vim
After this line, you place the actual contents of the document. Typically, this starts with a longer description of what this document is about, and what it will explain. This could include your name and contact info (email address, homepage etc.). After that, a table of contents could be added if the document is long enough to need one (we add it here as an example). It could in the case of myscript.vim look like this (including the first line): *myscript.txt* Documentation for example script myscript.vim Script : myscript.vim – Example script for vim developers Author : Kim Schulz Email : Changed: 01/01/2007 =========================================================== * myscript-intro* 1. Overview~ This document gives a short introduction to the example script myscript.vim. This script is made as an example for vim users on how to structure a simple vim plugin script such that it is easy to read and figure out. The following is covered in this document: 1. Overview |myscript-intro| 2. Mappings |myscript-mappings| 3. Functions |myscript-functions| 4. Todo |myscript-todo| =========================================================== [ 179 ]
Vim Scripting
In this example, we use most of the formatting tags you have available for Vim documents. Let's go through them one by one to see what they mean. The first markup is the *...* which marks keywords that the Vim help system can jump to. In this case, we have *myscript-intro*, which makes it possible to jump directly to this section in the documentation with: :help 'myscript-intro'
The next quite simple marker is the ~ after the Overview headline. This marks the line with a different color from the rest of the text. Then we come to the |...| around some keywords next to each item in the table of contents. This creates a link to a particular section with a matching keyword (marked with *...*). The lines with equal signs are just a good way to mark section borders; they are not actually a markup type. The following sections would be formatted in the same way, except that if they had a part that shows a piece of Vim code, then another markup would be used. An example of this could be in the function section. So let's take that as an example: =========================================================== *myscript-functions* 3. Functions~ Besides the functions available via mappings (as described in |myscript-mappings|) there are some extra global functions available. MyglobalfunctionB()~ This function is one of the global functions in this script. An example of usage could be: > :call MyglobalfunctionB() < Vim returns: Hello from the global-scope function myglobalfunctionB~ MyglobalfunctionC()~ This function is a global function that also calls one of the internal functions ("s:MyfunctionA()") in the script. An example of usage could be: > :call MyglobalfunctionC() < Vim returns: Hello from MyglobalfuncionC() now calling locally:~ This is the script-scope function MyfunctionA speaking~ =============================================================== [ 180 ]
Chapter 6
The special markup in this section is the >...< around the code examples. This is used to mark the code, while we use the ~ colored lines to mark the return from Vim. That is basically all you need to know in order to create readable documentation for your script. When the user wants to install the documentation, he or she places it in VIMHOME/doc/ and then he or she uses the following command: :helptags docdir
where docdir is the path to VIMHOME/doc/. If any of the keywords you have added is already used, then Vim will give you a warning, and you have to change it before distributing the documentation. Want to distribute your documentation in multiple languages? Take a look in the help system for more information: :help 'help-translated'
Using External Interpreters
Even though you can do nearly everything with Vim scripts, there are, however, some things that might be smarter or faster to do in other languages. The developers of Vim have recognized this need, and therefore implemented the possibility to interface with other scripting languages from within Vim. There are in particular three languages that you have access to: •
Perl
•
Python
•
Ruby
In the following sections, will we take a peek at how to interface with these scripting languages and which variables you have access to. The support for these language interfaces is not included in Vim by default, and you will have to either compile Vim yourself to get it, or find a precompiled version that supports them. To check if your version supports one of the languages, you simply run Vim on the command line with -version argument: vim –version
[ 181 ]
Vim Scripting
Then, you look through the list of features to see if it has one of the following in the list +perl +python +ruby
It needs to say + in front of the language name to show that it is included. If it instead says e.g. -perl, then Perl support is not included. Alternatively you can just open up Vim and then test for the features with the has() function: :echo has("perl") :echo has("python") :echo has("ruby")
It should return 1 for the languages you have support for.
Vim Scripting in Perl
Perl is a very popular scripting language that has been around for quite some time now. It is very powerful when it comes to parsing text and other similar tasks. This also makes it very useful from within Vim. The simplest way to use Perl in Vim is with the following command: :perl command
This executes the Perl command in the argument command. Note that the values you set with the Perl command will persist throughout the entire Vim editing session. Often you would, however, like to execute more than just a single command and hence you have to use another command. In that case you can use: :perl Set(100, "fooo") Replace line 100 in current buffer with
new text.
•
$curwin->SetCursor(15,8)
Set cursor at line 15, column 8 in current window.
You can find a full list of the Vim-specific functions you can use from within Perl by looking in the help system with: :help perl-pverview
If you put Perl code in your script, you should always remember to check if the user has support for Perl in his or her version of Vim. It is always a good idea to have your Perl code wrapped in Vim functions in your script. This way it is easy to implement your script, and for an inexperienced user, the script will look normal and work as usual. An example of how to wrap Perl in a function could be: function MoveCursor(row,col) if has("perl") perl Cursor(); VIM::Msg("Old position was: ($oldrow,$oldcol)"); $curwin->Cursor(row,col); [ 183 ]
Vim Scripting EOF else echo "perl not available. canceling function call" endif endfunction
This function gets the old position of the cursor in the current window, prints that position, and then moves the cursor to the position that matches the two arguments for the function (row and column). If the user does not have Perl support, then a message about this will be written. Note how the EOF is placed entirely to the left, even though the rest of the code is indented. This is strictly needed in order for Vim to be able to recognize it as the endpattern.
Vim Scripting in Python
Through the recent years, Python has become the favorite scripting language for many programmers. This is mainly for its ease of use and strict rules about indenting (which lead to more readable code). As with Perl, there is also an interface for Python in Vim, such that you can break out of Vim and use Python in your script. There are three main ways to use Python. When you only want to execute a single Python statement from Vim: :python statement
An example could be : :python print "hello Vim developer"
If you want to execute a larger amount of Python code at the same time, you can use the following from Vim: :python